Uteroglobin (UG) is an antiinflammatory protein secreted by the epithelial lining of all organs communicating with the external environment. We reported previously that UG-knockout mice manifest exaggerated inflammatory response to allergen, characterized by increased eotaxin and Th2 cytokine gene expression, and eosinophil infiltration in the lungs. In this study, we uncovered that the airway epithelia of these mice also express high levels of cyclooxygenase (COX)-2, a key enzyme for the production of proinflammatory lipid mediators, and the bronchoalveolar lavage fluid (BALF) contain elevated levels of prostaglandin D2. These effects are abrogated by recombinant UG treatment. Although it has been reported that prostaglandin D2 mediates allergic inflammation via its receptor, DP, neither the molecular mechanism(s) of DP signaling nor the mechanism by which UG suppresses DP-mediated inflammatory response are clearly understood. Here we report that DP signaling is mediated via p38 mitogen–activated protein kinase, p44/42 mitogen–activated protein kinase, and protein kinase C pathways in a cell type–specific manner leading to nuclear factor–κB activation stimulating COX-2 gene expression. Further, we found that recombinant UG blocks DP-mediated nuclear factor–κB activation and suppresses COX-2 gene expression. We propose that UG is an essential component of a novel innate homeostatic mechanism in the mammalian airways to repress allergen-induced inflammatory responses.
Introduction
Allergic inflammatory disorders result from a complex interplay of genetic and environmental factors. Even though the incidence of these disorders is on the rise throughout the industrialized world, the majority of the population manages to avert allergic inflammatory diseases. Thus, it appears that homeostatic mechanisms may exist to repress allergen-mediated inflammatory responses.
Uteroglobin (UG) is a steroid-inducible, low molecular weight, multifunctional, secreted protein with antiinflammatory/antichemotactic properties (for review see reference 1). It is the founding member of a newly designated Secretoglobin superfamily of proteins (2), constitutively expressed by the epithelial lining of all organs that communicate with the external environment. Structurally, UG is a homodimer in which the 70-amino acid subunits are covalently linked in antiparallel orientation by two interchain disulfide bridges forming a central hydrophobic cavity (3, 4). It has been suggested that this central cavity is capable of binding hydrophobic ligands such as steroid hormones, polychlorinated biphenyls, and retinol. The physiological importance of this property of UG remains unclear.
We reported previously that OVA immunization and challenge of UG-KO mice (5) cause highly exaggerated allergic response characterized by increased expression of Th2 cytokines, eotaxin, and eosinophil infiltration in the lungs (6). This suggests that UG may have a protective role against allergen-induced inflammatory response. However, the molecular mechanism(s) by which UG prevents allergen-induced inflammatory response remains unclear.
It is generally accepted that mast cells and other cell types in the lungs orchestrate allergic inflammation through the production of various cytokines and lipid mediators of inflammation such as prostaglandin D2 (PGD2) (7–11). Previous studies have shown that antigen/allergen challenge stimulates the release of arachidonic acid (AA) from membrane phospholipids by cytosolic phospholipase A2 catalysis (12). Moreover, it has been reported that PGD2 reinforces low dose antigen-stimulated Th2-type inflammatory responses via expression of macrophage-derived chemokines (13). More recently, it has been demonstrated that targeted overexpression of PGD2 synthase gene in the lungs of transgenic mice causes elevated levels of Th2 cytokines and eosinophil infiltration (14). Prostaglandins are produced from oxidation of arachidonic acid by cyclooxygenases (COX), also known as prostaglandin H synthase (15). Between the two isoforms of COX, COX-1 is constitutively expressed, whereas COX-2 is agonist inducible and plays critical roles in the production of proinflammatory lipid mediators (for review see reference 16). Moreover, it has been reported that: (a) the expression of COX-2, a critical enzyme for the production of proinflammatory lipid mediators in the respiratory system, is elevated in patients with allergic inflammation and asthma (17); (b) bronchoalveolar lavage fluids (BALFs), after acute antigen challenge (7) and those from patients with allergic asthma, contain high levels of PGD2; (c) mice in which PGD2 receptor, DP, is disrupted by gene targeting are refractory to developing allergen-induced airway inflammation (18); and (d) treatment of mice with a synthetic DP antagonist prevents allergic inflammation (19). Together, these results suggest that allergen-induced activation stimulates the release of PGD2, which mediates inflammation via DP signaling. However, the molecular mechanism(s) of DP signaling leading to allergic inflammatory response and the mechanism(s) by which UG represses allergen-induced inflammatory responses, until now, remained obscure.
To understand the molecular mechanism(s) of the exaggerated allergic response in UG-KO mice, we first performed detailed analyses of inflammatory response in OVA-sensitized and OVA-challenged UG-KO mice. We uncovered that in addition to the elevated expression of eotaxin, Th2 cytokines, and pulmonary eosinophil infiltration reported previously (6), the UG-KO mice also manifest elevated levels of PGD2 in the BALF and show increased COX-2 gene expression in the lungs. More significantly, treatment of these mice with purified recombinant UG (rUG) markedly inhibits these responses. Using representative cell lines from the mammalian respiratory system, we conducted experiments to understand: (a) the molecular mechanism(s) of DP signaling and (b) the mechanism by which UG suppresses DP-mediated inflammatory response. The results of these studies show that DP signaling is mediated via p38 mitogen-activated protein kinase (MAPK), p44/42 MAPK, and protein kinase C (PKC) pathways in a cell type–specific manner leading to the activation of nuclear factor (NF)–κB. Activation of NF-κB then stimulates the expression of COX-2, an agonist (allergen)-inducible enzyme that is critical for the production of inflammatory lipid mediators. Most importantly, we found that UG binds PGD2, blocks DP signaling, inhibits NF-κB activation, and consequently suppresses COX-2 gene expression. Our results, for the first time, demonstrate that at least one of the mechanisms by which DP signaling mediates allergic inflammatory response is by activating NF-κB that stimulates COX-2 expression. We propose that UG is a critical component of an innate homeostatic mechanism to prevent inadvertent stimulation of allergen-induced, DP-mediated inflammatory response.
Materials And Methods
Cell Lines and Mice.
The human bronchial smooth muscle (BSM-2146) and human alveolar type II (A549) cells were from Clonetics-BioWhittaker, Inc. The NIH-3T3 cells were from American Type Culture Collection. UG-KO mice were generated as described previously (5). Both UG-KO and WT mice were maintained under germ-free conditions, and all experiments were performed according to an institutionally approved animal use protocol.
Production and Purification of rUG.
Sensitization and Antigen Challenge of WT and UG-KO Mice.
Lung Lavage, Tissue Fixation, Histology, and Immunofluorescence.
The mice were killed 24 h after the OVA challenge, and BALF was collected in 1 ml of PBS as reported previously (6, 14). Total leukocyte counts and numbers of each cell type per milliliter of BALF were determined as reported (6, 14). Lung tissues were fixed in 4% paraformaldehyde and embedded in paraffin blocks. Tissue sections were stained with Verfuerth-Luna by American Histolabs, Inc. The stained slides were examined by using a Carl Zeiss MicroImaging, Inc. Axiophot Microscope, and digital photomicrographs were recorded.
RNA Isolation and Northern Blot Analyses.
Total RNAs from whole lungs were isolated using either Trizol (Invitrogen) or RNAzol B (Teltest Inc.) following the suppliers' protocol. Total RNA (25–40 μg) loaded in each lane were resolved by electrophoresis on 1.5% formaldehyde-agarose gels, and RNAs were transferred to Hybond N+ (Amersham Biosciences), cross-linked, and hybridized with [α32P]-dCTP-labeled cDNA probes at 68°C. The probes were generated by RT-PCR using gene-specific primers (Table S1) and sequence was conformed by DNA sequencing. RNA loading was standardized by hybridizing the blots with GAPDH or β-actin probes (5).
Cytokine and Eotaxin-mRNA Determination by Real-Time Quantitative RT-PCR Analyses.
The total RNA (1 μg) from the lungs was reverse transcribed using SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen). Real-time quantitative RT-PCR was performed using ABI Prism 7000 Sequence Detection System (Applied Biosystems) with cDNA equivalent to 40 ng of total RNA and SYBR Green PCR Master Mix (Applied Biosystems) following the manufacturer's protocol. The following PCR conditions were used: denaturation at 95°C for 10 min followed by 50 cycles of amplification at 94°C for 15 s and 60°C for 1 min. The primers used for IL-4, IL-5, IL-13, eotaxin, and β-actin are presented in Table S1. The data from each PCR run was analyzed using ABI Prism Software version 1.01 (Applied Biosystems). The final data were normalized to β-actin and presented as fold induction compared with the expression level in WT mice. Quantitation was performed using three independent total RNA samples for each treatment group.
Semiquantitative RT-PCR Analyses for DP- and UG-mRNA Expression.
The total RNA (1 μg) was reverse transcribed, and semiquantitative PCR was performed to amplify target fragments of cDNA with 5 μl of reverse transcription product using human and mouse DP- and UG-specific sense and antisense primers (Table S2). The primers for GAPDH are as reported previously (6). The following PCR conditions were used: denaturation at 95°C for 1 min followed by 30 cycles of amplification at 95°C for 30 s, 60°C for 50 s, 68°C for 3 min, and a final incubation at 68°C for 10 min.
Cell Culture and Treatment.
BSM-2146 cells were grown in their specified SmGM-2 medium (Clonetics). The NIH-3T3 cells were grown in DMEM at 37°C with 5% CO2 as described above. Cells were gown to 70–80% confluence before treating with various effectors (PGD2, rUG, kinase inhibitors, etc.), washed once with OptiMem-1 containing 2.5% FBS and then treated with the indicated effectors for the indicated period at 37°C with 5% CO2. The extracellular signal-regulated kinase pathway–specific inhibitor (PD98059) and the p38 MAPK pathway–specific inhibitor (SB203580) were dissolved in DMSO and then added to the OptiMEM-1 to obtain the indicated final concentration. Protein kinase C inhibitor, bisindolyl maleimide III (BMIII), was dissolved in water and added to the medium to obtain the desired final concentrations. Cells were pretreated for 15 min with these inhibitors or with DMSO alone (control) before addition of PGD2 (up to 50 μM) or as indicated. BSM-2146 cells were incubated for 3 h, and NIH-3T3 cells were incubated for 2 h at 37°C with 5% CO2.
3H-PGD2–binding Assay.
Cells were grown to 70–80% confluence in their recommended growth medium in 6-well cell culture plates, washed once with OptiMEM-1 without serum, and then incubated with 10 nM 3H-PGD2 in the absence and presence of varying concentrations (1 nM–5 μM) of PGD2 for 30 min at 4°C in 3 ml of OptiMEM without serum. Unbound 3H-PGD2 was removed by washing the cells three times with PBS. Cells were then lysed with lysis buffer, and 300 μl of the cell lysate was applied on a strip of 3-mm Whatman filter paper in a scintillation vial, dried at room temperature, and counted in a scintillation counter.
Transfection of Cells with DP Antisense Oligonucleotides.
A phosphorothioate DP antisense oligo (5′-GTGACGGCATTGCACGAGAAGG-3′) was designed to inhibit DP expression. As a control, the corresponding sense-oligo (5′-CCTTCTCGTGCAATGCCGTCAC-3′) was used. When the cells reached 30–40% confluence, they were rinsed once with serum-free Opti-MEM I medium. The oligonucleotides were then delivered to the cells by Oligofectamine™ Reagent (Invitrogen) according to the manufacturer's instructions. After 28 h, the cells were stimulated with 1 μM PGD2, and total proteins were isolated. The relative levels of COX-2 protein were estimated by Western blot analysis.
Western Blot Analyses.
For detection of the phosphorylated proteins, cell lysates were prepared in presence of protease and protein phosphatase inhibitors. Total protein (40 μg) from each sample was resolved by electrophoresis using 7.5% or 12% SDS-polyacrylamide gels under reducing conditions and electrotransferred to polyvinylidene fluoride membrane (Immobilon P; Millipore Corporation). Immunoblot analysis was performed using either goat or rabbit polyclonal anti–COX-2 (Santa Cruz Biotech Inc.)/anti–phospho p38-MAPK/anti–phospho p44/42 antibodies (Cell Signaling), respectively. Horseradish peroxidase–conjugated anti–goat or anti–rabbit IgGs were used as the second antibody. Chemiluminescent detection was performed by using ECL system (Amersham Biosciences) according to the manufacturer's protocol. Total protein loading standard, β-actin, was detected using Actin (Ab-1) Kit (Oncogene Research Products) according to the protocol of the supplier.
Blocking DP-mediated COX-2 Expression by DP Monoclonal Antibody Treatment of the Cells.
The NIH-3T3 cells were grown to 70–80% confluence. The cells were washed three times with serum-free OptiMEM and incubated with DP monoclonal antibody (Abcam, UK) in varying dilutions at room temperature with gentle agitation. After washing the cells with serum-free OptiMEM, the cells were stimulated with 1 μM PGD2 for 2 h. Total RNA was isolated and analyzed for COX-2 mRNA expression by real-time quantitative RT-PCR (22) using the mouse COX-2–specific primers (Table S2). Fold COX-2 expression levels were normalized to18S rRNA (for 18S primers see Table S2) and compared with the average values in untreated control cells. All determinations were performed in triplicates, and each experiment was repeated at least twice. COX-2 protein expression was determined by Western blot analysis.
Nuclear Extract Preparation and Electrophoretic Mobility Shift Assay.
Cells were treated with PGD2 (50 μM) in the presence and absence of rUG (150 nM). Nuclear extracts were prepared using commercially available buffers (GENEKA Biotech Inc.) using supplier's protocol. Electrophoretic mobility shift analyses (EMSAs) were performed using the nuclear extracts (20 μg protein) on a nondenaturing 5% polyacrylamide gel with the following oligonucleotides: for human cell extracts, hNFκBwt, 5′-AGTTGAGGGGACTTTCCCAGGC-3′; hNFκBMu, 5′-AGTTGAGATGACT-CGCCCAGGC-3′; for mouse cell extracts, mNFκBwt, 5′-GAGGGTGAGGGGATTCCCTTAGTTAGGAC-3′; mNFκBmu, 5′-GAGAGGTGAGATGATAGCCTTAGTTAGGAC-3′; mNF-IL6, 5′-CTGCCGCTGCGGTTCTTGCGCAACTCACT-3′; mAP1, 5′-AAAGACAGAGTCACCACTAATT-3′. NF-κB double-stranded oligonucleotides were generated by annealing sense and antisense oligonucleotides without and with mutations (bold). Specificity of protein–DNA complexes was verified either by competing with cold NF-κBwt/NF-κBMu oligos or by immunoreactivity with polyclonal antibodies specific for p65/p50 subunits of NF-κB.
Statistical Analysis.
Data were analyzed by using GraphPad Prism software (version 3.03). Student's t test was used to determine the significant differences between the means; p-values of <0.05 were considered as statistically significant.
Results
UG Inhibits Th2 Cytokine Expression in Allergen-sensitized and -challenged UG-KO Mice.
We reported previously that UG-KO mice, sensitized and challenged with OVA, manifest exaggerated allergic response characterized by elevated eotaxin and Th2 cytokine expression and an increased number of eosinophils in BALF (6). In this study, we examined the effects of rUG on these parameters. We first treated the OVA-sensitized UG-KO mice with rUG before OVA challenge and determined the expression of eotaxin and representative Th2 cytokine mRNAs by real-time quantitative RT-PCR. Our results confirm that OVA sensitization and challenge increase the level of these mRNAs in the lungs of UG-KO mice (Fig. 1 A) as reported previously (6). More importantly, we demonstrate that rUG treatment of OVA-sensitized UG-KO mice before OVA challenge inhibits the expression of IL-4 (Fig. 1 A, i), IL-5 (Fig. 1 A, ii), IL-13 (Fig. 1 A, iii), and eotaxin (Fig. 1 A, iv), all of which are recognized markers of allergic inflammatory response. The reduced level of inflammatory response in the lungs of OVA-sensitized and challenged WT mice may be due to the stimulation of UG expression in response to increased levels of Th2 cytokines. This hypothesis is at least partially supported by the fact that IL-13 treatment stimulates UG production in rat lungs (23). Therefore, we analyzed UG mRNA and UG protein in the lungs of OVA-sensitized and challenged WT mice by semiquantitative RT-PCR and immunofluorescence, respectively. The results show that OVA sensitization and challenge of WT littermates of UG-KO mice expressed higher levels of UG mRNA (Fig. 1 B) and UG protein (Fig. 1 C) compared with their nonsensitized and nonchallenged counterparts. These results raise the possibility that the observed OVA-induced elevation of UG production in WT mouse lungs is likely to be a homeostatic response to counteract inadvertent stimulation of allergen-induced inflammation.
UG Inhibits Eosinophil Infiltration.
We further determined the effects of rUG treatment on the infiltration of various leukocytes including eosinophils in the BALFs of these mice. Quantitation of each cell type in the BALF show that although leukocytes were virtually undetectable in control WT and UG-KO mice (Fig. 1 D); OVA sensitization and challenge caused an upward shift not only in the total number of infiltrated cells but also in the cell types. Most importantly, rUG treatment of OVA-sensitized UG-KO mice before OVA challenge significantly inhibited the number of eosinophils in BALF (Fig. 1 D). These results demonstrate that UG is a potent inhibitor of allergen-induced lung infiltration of leukocytes, especially eosinophils. Histological analysis of the lungs of OVA-sensitized and OVA-challenged UG-KO and WT mice for eosinophil infiltration show that eosinophils in the lungs of control WT and UG-KO mice are barely detectable (Fig. 1 E, frames 1 and 2). In contrast, although only a slight increase in the level of eosinophil infiltration was detectable in the lungs of OVA-sensitized and -challenged WT mice (Fig. 1 E, frame 3), it was markedly elevated in those of the UG-KO mice (Fig. 1 E, frame 4). Most importantly, when the OVA-sensitized UG-KO mice were treated with rUG before OVA challenge, it markedly inhibited eosinophil infiltration in the lungs (Fig. 1 E, frame 5). These results demonstrate that UG is a potent inhibitor of allergen-induced eosinophil infiltration in the lungs.
UG Inhibits Allergen-Induced Elevation of PGD2 Levels in BALF.
Elevated levels of PGD2 in the BALF of patients with allergic asthma have been reported, and it has been suggested that PGD2 is a potent chemoattractant for eosinophils (11, 24). Thus, we sought to determine the level of PGD2 in the BALF of OVA-sensitized and -challenged UG-KO and WT mice. In addition, we studied the effects of rUG treatment on PGD2 levels in BALF of OVA-sensitized and challenged mice. The results (Table S3) show that a very low level of PGD2 is detectable in the BALF of both UG-KO and WT control mice. In contrast, although PGD2 levels in BALF of OVA-sensitized and challenged WT littermates increased moderately from the baseline, higher levels of PGD2 were detected in those of the OVA-sensitized and challenged UG-KO mice. Most importantly, PGD2 levels in the BALF of OVA-sensitized UG-KO mice that were treated with rUG before OVA challenge were significantly reduced compared with those without rUG pretreatment (Table S2). These results suggest that UG inhibits allergen-induced elevation of PGD2 levels in the lungs.
UG Inhibits Allergen-induced Stimulation of COX-2 Gene Expression In Vivo.
COX-2 gene expression in the lungs of allergic asthma patients is markedly elevated (17). Since COX-2 is an agonist-inducible enzyme that catalyzes the production of proinflammatory lipid mediators, we first analyzed COX-2 gene expression in the lungs of nonsensitized and unchallenged UG-KO and WT littermates. The results show that a low level of COX-2 mRNA expression is detectable in the lungs of the WT littermates (Fig. 2 A, lane 1) and in those of the UG-KO (Fig. 2 A, lane 2) mice. In contrast, although COX-2 mRNA expression in the lungs of OVA-sensitized and challenged WT mice increased slightly (Fig. 2 A, lane 3), the expression levels in UG-KO mice were markedly elevated (Fig. 2 A, lane 4). Most importantly, treatment of OVA-sensitized UG-KO mice with rUG before OVA challenge showed an appreciable inhibition of COX-2 mRNA expression (Fig. 2 A, lane 5). We also found that compared with the intensity of COX-2 immunofluorescence in the bronchioles of control WT and UG-KO mice (Fig. 2 B, frames 1 and 2) immunofluorescence of the OVA-sensitized and challenged WT mice were only slightly elevated (Fig. 2 B, frame 3). However, COX-2 immunofluorescence was strikingly intense in the bronchioles of OVA-sensitized and challenged UG-KO mice (Fig. 2 B, frame 4). Significantly, the intensity of COX-2 immunofluorescence in OVA-sensitized UG-KO mice, treated with rUG before OVA challenge was markedly less intense (Fig. 2 B, frame 5), suggesting that UG suppresses allergen-induced stimulation of COX-2 gene expression in the lungs.
PGD2 Stimulates COX-2 Gene Expression In Vitro.
It has been reported that acute antigen challenge leads to increased PGD2 levels in human airways (7). Moreover, allergen-induced inflammation in asthmatic airways is associated with high levels COX-2 expression (17). Further, in DP-null mice allergen challenge do not elicit airway inflammation (18). The results of our present study show elevated levels of PGD2 in the BALF (Table S3) and high level of COX-2 gene expression in the lungs of OVA-sensitized and challenged UG-KO mice (Fig 2, A and B). Thus, we sought to determine if there is a link between the elevated levels of PGD2 and COX-2 gene expression. We used three representative cell types (i.e., epithelial, smooth muscle, and fibroblast cells) of the mammalian respiratory system. These cell types were used because these cell types, in cooperation with the immune system, orchestrate the initiation and propagation of allergen-induced inflammatory responses (25–27). We treated the cells with varying concentrations of PGD2 and determined COX-2 mRNA and protein expression by Northern and Western blot analyses, respectively. The results show that PGD2 stimulates COX-2 mRNA expression and that this stimulation is time (Fig. 2, C and D) and dose dependent (Fig. 2, E and F). The results of Western blot analyses show that PGD2 stimulates COX-2 protein expression in both NIH-3T3 and BSM-2146 cells (Fig. 2, G and H). Consistent with these results, we also found that PGD2 stimulates COX-2 mRNA expression in A549 cells (not depicted). Treatment of the cells with ethanol, used as solvent for PGD2, did not stimulate COX-2 expression. Together, our results indicate that PGD2 acts as a potent stimulant for COX-2 expression in three cell types of the respiratory system.
PGD2-Stimulated COX-2 Expression is Mediated via DP.
It has been demonstrated that PGD2 mediates allergen-induced airway inflammation via its receptor, DP (18). In this study, we found that PGD2 stimulates COX-2 expression in three different cell types of the respiratory system. To understand the mechanism(s) of PGD2-stimulated COX-2 gene expression and inflammatory lipid mediator production, we first determined DP expression and PGD2 binding using BSM-2146 and NIH-3T3 cells. The results of semiquantitative RT-PCR analyses show that these cells express DP-mRNA (Fig. 3 A, inset). We also performed competition-binding assays using 3H-PGD2 as the ligand. The results show that 3H-PGD2 binding on these cells is saturable, and it is displaced by nonradioactive PGD2 in a dose-dependent manner (Fig. 3 A). Since it has been reported that a chemoattractant receptor, CRTH2, also binds PGD2 in some cell types and may transduce some of the biological effects of this eicosanoid, we tested the expression of CRTH2-mRNA in these cells. The results show that the CRTH2-mRNA level is virtually undetectable (not depicted), indicating that CRTH2 is unlikely to be the predominant pathway for PGD2 signaling in these cell types.
To confirm whether PGD2-stimulated COX-2 expression is mediated via DP, we treated the NIH-3T3 cells with DP monoclonal antibody and stimulated these cells with PGD2. We analyzed both COX-2 mRNA and COX-2 protein expression by real-time quantitative RT-PCR and Western blot analyses, respectively. The results show that although a low level of COX-2 mRNA expression is detectable in PGD2-untreated control cells (Fig. 3 B, bar 1), PGD2 treatment stimulates markedly higher levels of COX-2 expression (Fig. 3 B, bar 2). In contrast, COX-2 expression in cells treated with three different dilutions of DP monoclonal antibody before PGD2 stimulation was significantly reduced (Fig. 3 B, bars 3–5). Consistent with these findings, the results of Western blot analysis show that the PGD2 untreated cells produced virtually no COX-2 protein (Fig. 3 C, top row, lane 1); however, those treated with PGD2 markedly stimulated COX-2 protein expression (Fig. 3 C, top row, lane 2). In contrast, cells treated with three different dilutions of DP monoclonal antibody showed a marked reduction in COX-2 protein in a dose-dependent manner (Fig. 3 C, top row, lanes 3–5). To further confirm that PGD2-stimulated COX-2 expression is DP mediated, we treated the cells with DP antisense s-oligonucleotide before PGD2 stimulation. The results show that although COX-2 protein expression in cells without PGD2 treatment is undetectable (Fig. 3 D, lane 1), in PGD2-treated cells (Fig. 3 D, lane 2) and in cells treated with sense s-oligo before PGD2 stimulation (Fig. 3 D, lane 3) it is readily detectable. Moreover, marked inhibition of COX-2 protein expression was observed in cells that are first treated with DP antisense s-oligo before PGD2 treatment (Fig. 3 D, lane 4). Together, these results provide strong evidence that the stimulation of COX-2 expression by PGD2 is mediated via DP.
Molecular Mechanism(s) of DP-Signaling is Cell Type Specific.
Although our results demonstrate that PGD2 stimulates COX-2 expression via DP, the molecular mechanisms of DP signaling, until now, remained unclear. Thus, we studied the mechanism(s) of DP signaling using our in vitro experimental system. The prostanoids signal through specific G protein–coupled receptors, and it has been reported that many agonists of these receptors activate the MAPK pathway (28). Thus, we analyzed the phosphorylation-dependent activation of kinase pathways. We treated the cells with PGD2 for varying lengths of time and analyzed tyrosine phosphorylation of MAPK by Western blot analysis using anti–phospho p38-MAPK (Tyr-182) and anti–phospho p44/42-MAPK (Tyr-204) antibodies. PGD2 stimulates the phosphorylation of both p38- (Fig. 4 A) and p44/42-MAPK (Fig. 4 B) in BSM-2146 cells, whereas in NIH-3T3 cells PGD2 stimulates the phosphorylation of p38 (Fig. 4 C) but not p44/42 MAPK (not depicted). These results indicate that DP signaling is mediated via the p38 and p44/42 MAPK pathways in a cell type–specific manner. We further confirmed the above findings by analyzing the effects of specific p38 MAPK inhibitor, SB203580, at varying concentrations, on PGD2-induced COX-2 expression in NIH-3T3 cells. The results show that SB203580 inhibits PGD2-induced COX-2 mRNA expression in a dose-dependent manner (Fig. 4 D). Most importantly, we found that rUG treatment inhibits PGD2-stimulated tyrosine phosphorylation of both p38 and p44/42 MAPK in BSM-2146 (Fig. 4 E) and p38 MAPK in NIH-3T3 cells (Fig. 4 F).
DP Signaling Is Mediated via p38, p44/42, and PKC-activating NF-κB and COX-2 Expression.
To determine the possible involvement of other kinase pathways in DP signaling, we treated these cells with PGD2 in the absence and presence of highly selective inhibitors of PKC (BMIII), p38 (SB203580), and p44/42 MAPK (PD98059), and determined COX-2 mRNA expression by Northern blot analysis. We found that in BSM-2146 cells the inhibitors of all three kinases suppressed PGD2-mediated stimulation of COX-2 mRNA expression at varying degrees (Fig. 5 A), suggesting that in these cells DP signaling is mediated via activation of p38 and p44/42 MAPK and PKC pathways. In contrast, PGD2-mediated stimulation of COX-2 expression in NIH-3T3 cells was inhibited only by p38 MAPK inhibitor, SB203580 (Fig. 5 B), whereas in A549 cells only the PKC-specific inhibitor (BMIII) suppressed such stimulation (not depicted).
Current evidence indicates that both MAPK and PKC are capable of activating NF-κB, a pivotal transcription factor that regulates many genes that are critical for inflammation and immunity (29). Therefore, we first examined the effects of specific inhibitors of NF-κB, pyrrolidine dithiocarbamate and NF-κB SN50, on PGD2-stimulated COX-2 mRNA expression. These agents are effective in inhibiting PGD2-mediated stimulation of COX-2 mRNA expression in both BSM-2146 (Fig. 5 C) and NIH-3T3 (Fig. 5 D) cells in a dose-dependent manner. These results suggest that DP-mediated COX-2 expression requires NF-κB activation. To determine whether the aforementioned kinase inhibitors inhibit DNA binding activity of NF-κB, we also conducted EMSAs to determine the effect of SB203580, PD98059, and BMIII on PGD2-stimulated NF-κB activation in both NIH-3T3 and BSM-2146 cells. The results show that although in BSM-2146 cells all three inhibitors suppress the binding of NF-κB probe (Fig. 5 E), in NIH3T3 cells only SB203580 was inhibitory (Fig. 5 F). Further, to confirm that NF-κB activation is required for the PGD2-stimulated COX-2 expression, we performed the COX-2 promoter-reporter (luciferase) activation assays by transfecting NIH-3T3 cells with reporter constructs, pGL2-COX2WT or pGL2-COX2MU (30), containing WT and mutant NF-κB binding site, respectively. Cells transfected with the vector alone served as controls. The results show that although PGD2 stimulates luciferase activity in the cells transfected with the pGL2-COX2WT construct, cells transfected with pGL-COX2MU construct had markedly reduced level of luciferase activity (Supplemental Materials and Methods and Fig. S1). Together, these results indicate that DP signaling is mediated via p38 and p44/42 MAPK and PKC pathways in a cell type–specific manner leading to NF-κB activation, required for COX-2 gene expression.
UG Inhibits PGD2-Stimulated COX-2 Gene Expression In Vitro.
We have shown that UG counteracts allergen-induced COX-2 gene expression in OVA-challenged allergic model of UG-KO mice (see the UG Inhibits Allergen-induced Stimulation…). Since OVA induces elevated levels of PGD2 in BALF, we determined if UG inhibits PGD2-mediated stimulation of COX-2 expression in vitro. We treated BSM-2146 and NIH-3T3 cells with PGD2 in the presence and absence of varying concentrations of UG and determined COX-2 mRNA expression by Northern blot analysis. We found that in both BSM-2146 (Fig. 6 A) and NIH-3T3 (Fig. 6 B) cells UG inhibits PGD2-induced COX-2 mRNA expression in a dose-dependent manner. Myoglobin (MG), a nonspecific control does not inhibit PGD2-stimulated COX-2 expression. Together, these results suggest that UG specifically inhibits PGD2-mediated COX-2 expression.
UG Inhibits COX-2 Expression by Suppressing PGD2-stimulated NF-κB Activation.
It has been reported that the COX-2 gene expression is regulated by at least three major nuclear factors (i.e., NF-κB, NF-IL6, and AP-1) located in the 5′ regulatory region of this gene. Since our results show that PGD2-stimulated COX-2 expression is mediated via NF-κB activation, we determined whether UG treatment of the cells blocks PGD2-mediated activation of this transcription factor. We performed EMSAs to identify and confirm the nuclear factor(s) specifically activated by PGD2 treatment of the cells leading to COX-2 expression. The results show that among the three major nuclear factors PGD2 specifically activates NF-κB (Fig. 6 C) in a time-dependent manner (Fig. 6 D). To determine whether UG blocks PGD2-induced NF-κB activation, we performed EMSA using nuclear extracts from NIH-3T3 and BSM-2146 cells treated with PGD2 in the presence and absence of UG. The results show that UG specifically inhibits DP-mediated activation of NF-κB in both BSM-2146 (Fig. 6 E) and NIH-3T3 (Fig. 6 F) cells. Our results provide strong evidence that UG inhibits PGD2 receptor–mediated inflammation by blocking the activation of NF-κB.
Discussion
Allergen-induced hypersensitivity and inflammation represent a growing public health problem in the industrialized world, affecting nearly 30% of the population <30 yr of age (31). The molecular mechanisms that underlie the pathophysiology of these conditions are poorly understood. In the present study, we uncovered several new and important aspects of allergen-induced inflammatory response and a possible innate homeostatic mechanism that represses such responses. More specifically, we found that: (a) in response to allergen (OVA) sensitization and challenge, UG-KO mice manifest significantly elevated levels of COX-2 gene expression in the lungs compared with that of the WT littermates; (b) in response to OVA sensitization and challenge, these mice also express elevated levels of eotaxin and Th2 cytokines, such as IL-4, IL-5, and IL-13, manifesting a markedly increased eosinophil infiltration in the lungs; and (c) most importantly, rUG treatment of the OVA-sensitized UG-KO mice before OVA challenge abrogates these allergic inflammatory responses. Using an in vitro system, we further demonstrate that (a) PGD2 mediates COX-2 expression specifically via its receptor, DP, and (b) DP signaling is mediated via p38 MAPK, p44/42 MAPK, or PKC in a cell type–specific manner activating NF-κB required for the expression of COX-2, an agonist-inducible enzyme critical for the production of inflammatory lipid mediators. Further, rUG treatment of the cells blocks PGD2-stimulated NF-κB activation and COX-2 expression.
PGD2 is an inflammatory eicosanoid released into human airways upon acute antigen challenge (7). Moreover, human lung and mast cells have been reported to generate PGD2 in a calcium-dependent manner (9, 11). Our results show a marked reduction of PGD2 levels in the BALF of OVA-sensitized UG-KO mice that are treated with rUG before OVA challenge. Previous reports indicate that UG is a potent antichemotactic agent for monocytes and neutrophils (32, 33). Thus, UG is likely to have similar antichemotactic effects on eosinophils and mast cells preventing lung infiltration. Since activated mast cells secrete PGD2, rUG may reduce the level of this eicosanoid in the lungs of allergen-challenged UG-KO mice by inhibiting the migration of mast cells. Recently, using a mouse model of allergic airway inflammation Honda et al. (13) have reported that inhalation of aerosolized PGD2 by OVA-sensitized mice precipitated allergic inflammatory response in the lungs. Since acute antigen challenge stimulates PGD2 release (7) and this prostanoid is a potent chemoattractant for human eosinophils (24), it is likely that UG gene is strategically expressed in the airway epithelia to protect this vital organ from allegen-induced, PGD2-mediated eosinophil infiltration.
Although it has been reported that PGD2 mediates allergic inflammation via its receptor, DP, the molecular mechanism(s) of DP signaling and how it mediates inflammatory response, until now, remained unclear. Since one of the major goals of this study was to understand the mechanism(s) by which UG prevents the allergen-induced inflammatory responses, we determined whether UG plays any role in down-regulating DP signaling and PGD2-stimulated COX-2 expression. We first uncovered that DP signaling is mediated via p38 MAPK, P44/42 MAPK, and PKC in a cell type–specific manner leading to the activation of NF-κB and subsequent stimulation of COX-2 gene expression. To our knowledge, this is the first report delineating the mechanism of DP-mediated activation of NF-κB and consequent stimulation of COX-2 gene expression in three different cell types. Most importantly, our results show that rUG inhibits DP-mediated NF-κB activation and OVA-induced stimulation of PGD2 levels in BALF. In addition, rUG inhibits OVA-induced expression of eotaxin, several Th2 cytokines (i.e., IL-4, IL-5, IL-13), and COX-2, a critical enzyme for the production of inflammatory lipid mediators. Cytokines such as IL-4, IL-5, and IL-13 are capable of inducing IgE production and activation of mast cells and eosinophils (34, 35). Our results show that the level of these cytokines is markedly reduced by rUG treatment of OVA-sensitized UG-KO mice before OVA challenge. IL-13 has been suggested to play a critical regulatory role in mediating allergic inflammation (36–38) and has been reported to stimulate UG gene expression in rat lungs (23). Therefore, UG may have a homeostatic role in down-regulating allergic inflammatory response. Our results indicate that UG suppresses allergic inflammatory response not only by inhibiting COX-2 expression and reduction of PGD2 levels in the BALF but also by inhibiting the production of proinflammatory cytokines and eotaxin.
Structural studies have delineated that UG is a homodimeric protein in which the identical 70-amino acid subunits are joined in antiparallel orientation by two interchain disulfide bonds forming a central hydrophobic cavity (3, 4). Previous reports have indicated that UG binds hydrophobic ligands such as progesterone (39) and retinol (40), raising the possibility that it may also bind and possibly sequester other hydrophobic ligands such as PGD2. We have conducted molecular modeling studies using the crystallographic coordinates of human rUG (24), and we have shown that the structural features of UG readily allow docking of PGD2 into its central hydrophobic cavity (Supplemental Materials and Methods and Fig. S2). Consistent with this finding, the results of our experiment show that when 3H-PGD2 is incubated with rUG a radioactive PGD2–UG complex is readily detectable by SDS-PAGE analysis (Supplemental Materials and Methods and Fig. S3). Thus, it appears that at least one of the mechanisms by which UG may prevent DP-mediated NF-κB activation and COX-2 gene expression is by sequestering this eicosanoid into its hydrophobic cavity, thereby suppressing the inflammatory response. Since UG is one of the most abundant proteins constitutively expressed in the lungs (μg/ml of BALF) and our results indicate that the maximum PGD2 levels in the BALF of OVA-sensitized and -challenged WT mice are pg/ml range, the stoichiometry appears to favor our proposed model.
PGD2 is a potent chemoattractant for eosinophils (24), and transgenic mice in which PGD2 synthase overexpression is targeted to the lungs manifest elevated Th2 cytokines and eosinophil infiltration (14). Recently, it has been reported that aerosolized PGD2 reinforces allergen-induced inflammatory response via macrophage-derived chemokines (13). Although our results reconfirm the important role of PGD2 in allergen-induced inflammatory response and the critical role of UG in suppressing such response mediated via DP signaling, these results also suggest that UG suppresses the production of OVA-induced chemokines. Our results also demonstrate that UG inhibits OVA-induced stimulation of eosinophil infiltration in the lungs of UG-KO mice. This observation is consistent with our previous finding that UG inhibits chemotaxis of monocytes and neutrophils (32, 33), although the molecular mechanism(s) of this effect of UG is not yet clear. Since UG binds to cell surface binding proteins (putative receptors) (41), future studies may delineate whether the antichemotactic effects of UG are mediated via these binding proteins.
It has been reported that although UG gene expression in the airway epithelia is constitutive, its expression is further augmented by glucocorticoids (42), the most effective drug currently available for the treatment of airway inflammation. Although the administration of nonsteroidal antiinflammatory drugs, which are known to inhibit both COX-1 and COX-2 enzymes (16), may have catastrophic consequences in ∼10% of the adults with asthma, it has been reported that selective COX-2 inhibitors such as rofecoxib and celecoxib do not manifest such adverse reactions (43, 44). Thus, the commonly available nonsteroidal antiinflammatory drugs may have effects in addition to the COX-2 inhibitory activity that causes the adverse complications in asthmatic patients. On the other hand, glucocorticoids, which inhibit COX-2 gene expression and reduce prostaglandin levels (45), remain the mainstay of current treatment for airway inflammation in asthma, which often has an allergic component. Although the mechanism(s) of antiinflammatory effects of glucocorticoids is still the subject of intense investigation, a recent study has elegantly demonstrated that targeted disruption of a glucocorticoid-inducible gene, lipocortin-1 (annexin-1) (46), in mice causes increased COX-2 expression and manifests an exaggerated inflammatory response (47). These results are consistent with those observed in our UG-KO mice.
We have reported previously that UG-deficient mice manifest abnormal deposition of fibronectin and IgA in the kidney glomeruli (48), reminiscent of human IgA nephropathy. Further, the results of our present investigation together with those of our previous study (6) indicate that UG-KO mice are susceptible to developing exaggerated inflammatory response to allergen challenge that is abrogated by rUG treatment. This may explain a long-standing unexplained linkage between inflammation in the respiratory system with disease pathology in the renal system. Several reports have documented that inflammatory lung diseases are associated with abnormal deposition of IgA and extracellular matrix proteins, such as fibronectin, in the renal glomeruli (49–53) as we have reported in UG-KO mice (5). These observations, together with the facts that (a) chronic airway inflammation markedly suppresses UG production (53) and (b) the UG-KO mice (5) develop IgA nephropathy (48) may further support the notion that pulmonary inflammation and renal pathologies are causally related. Our results demonstrate the importance of UG and perhaps of other proteins with similar properties in maintaining homeostasis in two different vital organs, the lungs and the kidneys.
The respiratory system and the renal system communicate with the external environment providing a portal of entry for microorganisms and other foreign antigens and allergens. During the last decade, tremendous progress has been made in delineating the innate defense mechanisms that exist against microbial pathogens (54). However, even after the pathogens are destroyed by the innate microbial defense system, the respiratory system must also cope with inadvertent stimulation of inflammation induced by antigens associated with these pathogens. The mechanism(s) by which the majority of the individuals avert inadvertent stimulation of inflammatory response despite constant exposure to foreign antigens remains a mystery. It is tempting to speculate that UG is strategically expressed in organs that communicate with the external environment and provide protection against allergen-induced stimulation of the inflammatory response. Further understanding of the molecular mechanisms of action of UG and other related proteins may allow the development of novel therapeutic approaches to at least some allergic and inflammatory disorders.
Acknowledgments
We thank Dr. Cherie A. Singer (University of Nevada School of Medicine, Reno, Nevada) for a generous gift of pGL2-COX-2WT and pGL2-COX-2MU luciferase constructs. We also thank Drs. A. Tager, I. Owens, S.W. Levin, and J.Y. Chou for critical review of the manuscript and valuable suggestions. The expert assistance of Shauna Everett and Rick Dreyfus with photomicrography is gratefully acknowledged.
References
A.K. Mandal, Z. Zhang, and R. Ray contributed equally to this work.
The online version of this article contains supplemental material.
A.K. Mandal's present address is Dept. of Medicine/Renal Unit, Harvard Medical School and Massachusetts General Hospital-East, Charlestown, MA 02129.
Abbreviations used in this paper: AA, arachidonic acid; BALF, bronchoalveolar lavage fluid; COX, cyclooxygenase; EMSA, electrophoretic mobility shift analysis; MAPK, mitogen-activated protein kinase; MG, myoglobin; OVA, ovalbumin; PGD2, prostaglandin D2; PKC, protein kinase C; rUG, recombinase UG; UG, uteroglobin.