METHODS

Cell culture

Normal human bronchial epithelial cells (cell line NL20) were obtained from ATCC and the human bronchial epithelial cell line BEAS-2B was provided by Berthold Fischer, Department of Internal Medicine II, University of Mainz. Both cells lines were established by transfection with the origin of replication defective SV40 large T plasmid. NL20 cells were cultured in Ham’s F12 medium with supplements according to the manufacturer’s protocol. BEAS-2B cells were cultured inRPMI supplemented with penicillin, streptomycin, and 10% fetal calf serum. Primary human bronchial epithelial cells were obtained from Biowhittaker and cultured according to the manufacturer’s protocol.

RT-PCR

Total RNA from NL20 and BEAS cells was isolated for RT-PCR with amplification of PPARγ and GAPDH cDNA as described previously.¹⁷

Gelatin zymography and ELISA

To examine the effect of PPARγ activators on MMP-9 gelatinolytic activity, cells were treated with TNF-α (10 ng/ml) or PMA (10 ng/ml) in the absence or presence of two different PPARγ activators—rosiglitazone (GlaxoSmithKline) and pioglitazone (Takeda)—in serum-free media for 12 hours. Culture supernatants were mixed in SDS-PAGE loading buffer (lacking reducing agents), applied to 10% SDS-polyacrylamide gels containing 1 mg/ml gelatin (Bio-Rad), and separated by electrophoresis. Subsequently, SDS was removed from the gels by two washes (15 minutes) with 2.5% Triton X-100 (VWR Scientific). After the washes, gels were incubated overnight (37°C) in zymography buffer (50 mmol/l Tris, pH 7.3, 10 mmol/l CaCl₂, and 0.05% Brij 35 (Sigma)) and stained with Coomassie brilliant blue (Sigma). Gelatinolytic activity was visualised as clear zones of lysis against a dark background. In some experiments cells were treated with TNF-α and PPARγ activators for 12 hours, then cultured again in regular media for 12 hours before additional stimulation with TNF-α in the absence or presence of rosiglitazone or pioglitazone.

TIMP-1 protein content in conditioned media was determined by ELISA according to the manufacturer’s instructions (R&D Systems).

Northern blot analysis

Total RNA (5 μg) was used for standard northern blotting.

Transient transfection assays

To examine whether PPARγ is functionally active in the cells used, NL20 and BEAS cells were transiently transfected with PPRE₃-TK-LUC²⁰ (provided by Dr Bruce Spiegelman, DFCI, Boston, MA, USA) and pCMV-β-GAL using superfect, according to the manufacturer’s protocol (Quiagen). Transfected cells were stimulated with rosiglitazone at the indicated concentrations. Cells were harvested after 24 hours and luciferase and β-galactosidase activity was measured using the Dual-Light assay (Tropix, Bedford, MA, USA).

To investigate the effect of PPARγ activators on NF-κB activity, BEAS cells were transiently transfected with a luciferase reporter construct containing three copies of the prototypic NF-κB site of the MHC I promoter and a pCMV.β-GAL construct as an internal control. Transfected cells were cultured for 16 hours before stimulation with TNF-α, with or without PPARγ activators. Cells were then harvested after 6 hours and lysates were subjected to luciferase and β-galactosidase assay (Tropix).

Electrophoretic mobility shift assay

For electrophoretic mobility shift assays (EMSA), human BEAS cells were stimulated for 2 hours with TNF-α (10 ng/ml) in the absence or presence of PPARγ activators before nuclear extracts were prepared. Standard EMSAs were performed, using an oligonucleotide (Genosys Biotechnologies, The Woodlands, TX, USA) spanning a prototypic κB site from the MHCI promoter.

Statistical analysis

The results of the experimental studies are reported as mean (SE). Differences were analysed by one way ANOVA followed by the appropriate post hoc test. A p value of <0.05 was regarded as significant.

RESULTS

Expression of PPARγ in human bronchial epithelial cells

RT-PCR was performed to establish the expression of PPARγ in human bronchial epithelial cells used in our experiments. NL20 cells and BEAS cells contained PPARγ mRNA as detected by a 376 bp RT-PCR product (fig 1A). The identity of the detected PPARγ bands was confirmed by DNA sequencing (data not shown).

Activity of PPARγ in human bronchial epithelial cells

To assess the presence of functional endogenous PPARγ we transiently transfected NL20 and BEAS cells with a PPRE-luciferase construct and stimulated cells with increasing amounts of the PPARγ activator rosiglitazone. Luciferase activity was assayed and normalised to β-galactosidase activity of a cotransfected pCMV-β GAL construct. Stimulation with rosiglitazone increased normalised luciferase activity in a concentration dependent manner in both cell lines (fig 1B and C). These results suggest the presence of inducible PPARγ activity in these cell types.

Inhibition of TNF-α or PMA induced MMP-9 gelatinolytic activity by PPARγ activators in human bronchial epithelial cells

Given the presence of functionally active PPARγ in bronchial epithelial cells, we examined the effect of PPARγ activators on MMP activity in these cells. Conditioned media harvested from unstimulated NL20 or BEAS cells exhibited little gelatinolytic activity as determined by substrate zymography. As expected, stimulation with TNF-α significantly induced MMP-9 activity in the supernatant but had no effect on MMP-2 (fig 2A). Incubation of cells with rosiglitazone or pioglitazone significantly reduced this increase in a concentration dependent manner with a maximal reduction to 60 (7)% with 10 μM pioglitazone or to 49 (8)% with 10 μM rosiglitazone (p<0.05 compared with TNF-α -stimulated cells, n=5) in NL20 cells, and to 60 (8)% with 10 μM pioglitazone or to 45 (9)% with 10 μM rosiglitazone (p<0.05 compared with TNF-α stimulated cells, n=3) in BEAS cells. None of the PPARγ activators significantly altered MMP-2 activity (fig 2A and B).

To examine whether the effects observed depended on the stimulus employed, we performed similar experiments using PMA as an inducer of MMP-9 activity. PMA treatment significantly induced MMP-9 activity, while concomitant treatment with rosiglitazone or pioglitazone diminished this increase significantly in a concentration dependent manner but had no effect on MMP-2 activity (fig 3A and B). Similar results were obtained with primary human bronchial epithelial cells; PMA significantly induced MMP-9 activity and both pioglitazone and rosiglitazone reduced this increase to 47 (3)% and 43 (1)%, respectively (fig 3C).

Sequential stimulation of cells with PPARγ activators revealed that cells do not become refractory to PPARγ stimulation in inhibiting TNF-α induced MMP-9 activity (fig 3D).

Effect of PPARγ activators on TIMP-1 expression in human bronchial epithelial cells

Since the effects of PPARγ activation on MMP expression did not differ in the two cell lines, further experiments were performed only in BEAS cells. To examine whether PPARγ activators shift the balance of matrix degrading enzymes and their inhibitors towards matrix stabilisation, we measured the release of TIMP-1 (the local inhibitor of MMP-9) after stimulation with PPARγ activators. As shown in fig 4, neither rosiglitazone nor pioglitazone had a significant effect on TIMP-1 protein content in the supernatants from BEAS cells.

Effect of PPARγ activators on MMP-9 mRNA expression in human bronchial epithelial cells

To investigate whether the decrease in MMP-9 activity by PPARγ activators resulted from reduced mRNA expression, northern blot analysis was performed after 6 hour stimulation. Both rosiglitazone and pioglitazone markedly reduced the TNF-α induced MMP-9 mRNA content in a concentration dependent manner, but did not affect mRNA levels of the constitutively expressed gene GAPDH (fig 5).

Effect of PPARγ activators on NF-κB activation in human bronchial epithelial cells

Previous work has shown that TNF-α induced MMP-9 expression in human bronchial epithelial cells is regulated by activation of NF-κB. To assess the effect of PPARγ activation on NF-κB, transient transfection experiments of a reporter construct containing three copies of the MHC class I NF-κB site were used. Stimulation of transfected cells with TNF-α (10 ng/ml) induced a 3.6 (0.6) fold increase in normalised luciferase activity which was reduced to 2.1 (0.2) fold by rosiglitazone and to 2.5 (0.4) fold by pioglitazone (fig 6A). To further investigate whether PPARγ activators inhibit binding of NF-κB transcription factors to the respective NF-κB DNA binding site, we performed electromobility shift assays using oligonucleotides corresponding to the MHCI NF-κB site. As shown in fig 6B, both rosiglitazone and pioglitazone inhibited TNF-α induced binding of the NF-κB proteins to DNA.

DISCUSSION

This study demonstrates the expression and functional activity of PPARγ in two different human bronchial epithelial cell lines and shows that PPARγ activators inhibit the expression of MMP-9 in these cells through an interaction of PPARγ with the MMP-9 inducing transcription factor NF-κB.

PPARγ expression was originally described in adipocytes and hepatocytes and recent work has demonstrated its presence in inflammatory cells, endothelial cells, and various cells in the intestine. Immunohistochemical staining of normal bronchial tissue revealed PPARγ immunoreactivity in epithelial cells,¹⁹ and in vitro experiments suggested the expression of this receptor in lung cancer cells²¹ and NIH A549 cells. Our study extends the knowledge about PPARγ in the respiratory system by showing PPARγ expression in two different cell lines of human normal bronchial epithelial cells. In addition, the study shows that PPARγ activators are capable of activating endogenous PPARγ in these cells by increasing the activity of a transfected PPRE-luciferase construct.

Stimulation of bronchial epithelial cells with two different PPARγ activators significantly attenuated cytokine induced gelatinolytic activity of MMP-9 but not of MMP-2, another matrix degrading enzyme which is also thought to be implicated in inflammatory airway processes. The ligands employed—rosiglitazone and pioglitazone—are most probably exhibiting their effects through the activation of PPARγ, given that these agents have high binding activities for PPARγ.^22,23 However, recent work has suggested that thiazolidinediones might exhibit PPARγ independent effects such as regulation of cell cycle processes in cells of the monocytic lineage.²⁴ We cannot therefore exclude the possibility that similar mechanisms might be involved in the effects investigated in this study.

The action of PPARγ activators on MMP-9 expression is independent of the inducing stimulus since these agents inhibited both TNF-α and PMA induced MMP-9 release, suggesting a direct effect on the expression of this gene. Transient transfection assays and gel shift analyses suggest that PPARγ exhibits its inhibitory effects on MMP-9 expression by limiting TNF-α induced NF-κB activation. In addition, PPARγ seems to limit binding of the NF-κB proteins to the respective DNA binding site. Previous work has shown inhibition of NF-κB activation by PPARγ activators in various cell types such as endothelial cells and monocyte/macrophages.^15,25 However, the underlying molecular mechanism of this interaction still needs to be elucidated.

Taken together, these data suggest a novel mechanism to counterbalance the release of matrix degrading MMP-9 from bronchial epithelial cells. PPARγ activation might therefore be useful for preserving the histological airway structure in chronic inflammatory processes, but in vivo experiments and clinical data are needed to determine the biological relevance of MMP-9 inhibition mediated by PPARγ activators.