Airway epithelial cell PPARγ modulates cigarette smoke-induced chemokine expression and emphysema susceptibility in mice

Abstract

Chronic obstructive pulmonary disease (COPD) is a highly prevalent, chronic inflammatory lung disease with limited existing therapeutic options. While modulation of peroxisome proliferator-activating receptor (PPAR)-γ activity can modify inflammatory responses in several models of lung injury, the relevance of the PPARG pathway in COPD pathogenesis has not been previously explored. Mice lacking Pparg specifically in airway epithelial cells displayed increased susceptibility to chronic cigarette smoke (CS)-induced emphysema, with excessive macrophage accumulation associated with increased expression of chemokines, Ccl5, Cxcl10, and Cxcl15. Conversely, treatment of mice with a pharmacological PPARγ activator attenuated Cxcl10 and Cxcl15 expression and macrophage accumulation in response to CS. In vitro, CS increased lung epithelial cell chemokine expression in a PPARγ activation-dependent fashion. The ability of PPARγ to regulate CS-induced chemokine expression in vitro was not specifically associated with peroxisome proliferator response element (PPRE)-mediated transactivation activity but was correlated with PPARγ-mediated transrepression of NF-κB activity. Pharmacological or genetic activation of PPARγ activity abrogated CS-dependent induction of NF-κB activity. Regulation of NF-κB activity involved direct PPARγ-NF-κB interaction and PPARγ-mediated effects on IKK activation, IκBα degradation, and nuclear translocation of p65. Our data indicate that PPARG represents a disease-relevant pathophysiological and pharmacological target in COPD. Its activation state likely contributes to NF-κB-dependent, CS-induced chemokine-mediated regulation of inflammatory cell accumulation.

Keywords: chronic obstructive pulmonary disease; lung inflammation; nuclear factor-kB; peroxisome proliferator-activating receptor-γ; rosiglitazone.

Fig. 1.

Airway epithelial cell peroxisome proliferator-activating receptor (Ppar)-g deficiency increases susceptibility to smoke-induced emphysema. Airway epithelial cell-specific, Pparg-targeted [knockout (KO)], or littermate control (CTL) mice were exposed to cigarette smoke [CS (SM)] for 6 mo. Age-matched mice were exposed to room air for 6 mo (NS). A: quantitative morphometry demonstrates a significant increase in CS-induced airspace enlargement (emphysema) in PPARγ-targeted mice. B: number of macrophages in smoke-exposed mice is presented, as defined by staining for Mac-3. Macrophage accumulation is increased in the lungs of conditionally targeted mice upon CS exposure. Data are shown as means ± SE (n = 4 mice/group).*P < 0.05 by t-test.

Fig. 2.

Loss of lung epithelial cell Pparg results in exaggerated chemokine expression. A: lung epithelial cell-specific Pparg targeted and littermate control mice were exposed to room air or CS for 6 mo. Whole lung tissue RNA was analyzed by quantitative real-time PCR (qPCR) for the genes indicated. B: Cxcl10 protein was measured in mouse whole lung tissue homogenates by ELISA. Data are shown as mean values ± SE (n = 4–7/group). #P < 0.05 by Mann Whitney U (MWU)-test.

Fig. 3.

Pharmacological activation of PPARγ using thiazolidinedione (TZD) limits smoke-induced inflammation. Wild-type (CTL) mice were exposed to room air or CS for 12 wk while fed a control diet or a diet containing 15 mg/kg rosiglitazone (RGZ) or 33 mg/kg pioglitazone (PIO). Lung inflammatory cell accumulation was assessed for all cells (A), macrophages (B), neutrophils (C), or lymphocytes (D). Data are shown as means + SE (n = 4–5/group). E: whole lung tissue RNA was analyzed by qPCR for the genes indicated. Data are shown as mean values (n = 4–7/group). *P < 0.05 by t-test. F: Cxcl10 protein was measured in whole lung tissue homogenates by ELISA. Data are shown as mean values ± SE (n = 4–6/group). #P < 0.05 by MWU test.

Fig. 4.

CS-induced lung epithelial cell chemokine expression is suppressed by activation of PPARγ. A and B: confluent cultures of lung epithelial cell lines were treated with cigarette smoke condensate (CSC, 75 μg) alone, or in combination with PPARγ-activating ligand, for 48 h. A: gene expression in mouse transformed Club-like (Clara) cells (mtCC) treated with CSC and RGZ (12 μM) was analyzed by qPCR for the genes indicated. B: H292 human lung airway-like epithelial cells treated with CSC, RGZ (12 μM), or Azelaoyl PAF (AzPAF, 1 μM) and analyzed for CCL5 and CXCL10 protein levels by xMAP. C and D: primary normal human bronchial epithelial (NHBE) cells were differentiated at the air-liquid interface (ALI) and treated with CSC (40 μg) and RGZ (12 μM) for 48 h. C: gene expression was analyzed by qPCR for the genes indicated. D: protein secretion into the basolateral medium was analyzed by xMAP. Data are shown as mean values ± SE. #P < 0.05 by MWU test. *P < 0.05 and **P < 0.09 by t-test.

Fig. 5.

PPARγ activation suppresses CSC-induced NF-κB transcriptional activity in human lung airway epithelial cells. PPARγ and NF-κB transcriptional activity was measured in human lung epithelial cells (H292), treated with RGZ and/or CSC for 48 h, using reporter assays. A: cells were transiently transfected for 12 h with (PPRE)3-TK-Luc and pCMV-β-gal, followed by treatment with RGZ (12 μM) and/or CSC (60 μg/ml), as indicated. Results are presented as mean relative luciferase units (RLU) normalized to β-gal activity. Treatment with RGZ or CSC increased PPARγ transcriptional activity. B: cells were transiently transfected for 12 h with pNF-(κB)5-E1B-Luc and pCMV-β-gal, followed by treatment with RGZ (12 μM) and/or CSC (60 μg), as indicated. Treatment with CSC increased NF-κB transcriptional activity, whereas RGZ decreased both baseline, and CSC-induced, NF-κB activity. Results are presented as mean RLU ± SE normalized to β-gal activity. *P < 0.05 by t-test.

Fig. 6.

Regulation of NF-κB activity by nonpharmacological PPARγ ligands or by ectopic PPARγ expression. A: human lung airway epithelial cells (H292) were transiently transfected with pNF-(κB)5-E1B-Luc and then treated with the AzPAF alone, or in combination with CSC (60 μg/ml). Cells were harvested and assayed for normalized luciferase activity (relative to β-Gal) after 48 h. Similar to RGZ, treatment with AzPAF decreased CSC-induced NF-κB transcriptional activation. B: H292 cells were transiently transfected with pNF-(κB)5-E1B-Luc and pSV40-RL along with control vector (V, empty pcDNA3.0), or vector containing a cDNA for wild-type (WT) human PPARγ (V + hPPARγWT). Cells were stimulated with CSC alone (60 μg/ml), or in combination with RGZ (6 μM), and assayed for normalized luciferase activity (relative to β-Gal) after 48 h. Ectopic expression of WT PPARγ suppressed CSC-induced NF-κB activity and potentiated the suppressive effects of RGZ. Results are presented as mean RLU ± SE normalized to β-gal activity. *P < 0.05 by t-test.

Fig. 7.

PPARγ binds NF-κB and regulates its activity at multiple levels. H292 cells were treated with CSC (60 μg/ml) and RGZ (12 μM), alone or in combination. Cells were harvested for nuclear, cytoplasmic extracts or whole cell extracts after 48 h unless otherwise noted. A: PPARγ and NF-κB complex was coimmunoprecipitated with antibodies against PPARγ, and blotted using p65 antibody. CSC and RGZ treatment promoted the association of PPARγ with NF-κB. B: cytoplasmic or nuclear extracts were assayed for NF-κB subunits of p65. CSC induced the nuclear translocation of p65. RGZ treatment inhibited the CSC-induced nuclear translocation of p65. C: cytoplasmic or nuclear extracts were assayed for p50. CSC did not induce the nuclear translocation of p50. RGZ treatment inhibited nuclear translocation of p50. D: cytoplasmic extracts were assayed for IκBα after 12 h. Whereas CSC induced degradation of IκBα, RGZ treatment inhibited the degradation to some extent. E: activation of PPARγ inhibits CSC-induced phosphorylation and activation of IKK. Total protein extracts were separated on 8% gel and immunoblotted using pIKKα/β (Ser¹⁷⁶)-specific antibody. β-Actin was used as a protein-loading control. Black line in the gel (A) represents that the lanes were run on the same gel but were noncontiguous.