Protective Role of Interleukin-10 in Ozone-Induced Pulmonary Inflammation

Background The mechanisms underlying ozone (O3)-induced pulmonary inflammation remain unclear. Interleukin-10 (IL-10) is an anti-inflammatory cytokine that is known to inhibit inflammatory mediators. Objectives We investigated the molecular mechanisms underlying interleuken-10 (IL-10)–mediated attenuation of O3-induced pulmonary inflammation in mice. Methods Il10-deficient (Il10−/−) and wild-type (Il10+/+) mice were exposed to 0.3 ppm O3 or filtered air for 24, 48, or 72 hr. Immediately after exposure, differential cell counts and total protein (a marker of lung permeability) were assessed from bronchoalveolar lavage fluid (BALF). mRNA and protein levels of cellular mediators were determined from lung homogenates. We also used global mRNA expression analyses of lung tissue with Ingenuity Pathway Analysis to identify patterns of gene expression through which IL-10 modifies O3-induced inflammation. Results Mean numbers of BALF polymorphonuclear leukocytes (PMNs) were significantly greater in Il10−/− mice than in Il10+/+ mice after exposure to O3 at all time points tested. O3-enhanced nuclear NF-κB translocation was elevated in the lungs of Il10−/− compared with Il10+/+ mice. Gene expression analyses revealed several IL-10–dependent and O3-dependent mediators, including macrophage inflammatory protein 2, cathepsin E, and serum amyloid A3. Conclusions Results indicate that IL-10 protects against O3-induced pulmonary neutrophilic inflammation and cell proliferation. Moreover, gene expression analyses identified three response pathways and several genetic targets through which IL-10 may modulate the innate and adaptive immune response. These novel mechanisms of protection against the pathogenesis of O3-induced pulmonary inflammation may also provide potential therapeutic targets to protect susceptible individuals.

Interleuken (IL)-10 is a pleiotropic cytokine that is produced by activated monocytes, macrophages, and helper T-cells, and B-cells. IL-10 reduces TNF-α and IL-6 production (Lang et al. 2002), and inhibits MIP-2 (Standiford et al. 1995). Previously, Reinhart et al. (1999) showed that intratracheal instillation of recombinant IL-10 in Sprague-Dawley rats before O 3 exposure significantly reduced O 3 -induced PMN infiltration and pulmonary hyperpermeability responses [fibronectin, albumin, bronchoalveolar lavage fluid (BALF) protein]. However, the relationship between IL-10 and other inflammatory mediators and downstream molecular events has not been investigated. IL-10 inhibits inflammation by suppressing macrophage CD86 expression leading to anergic T-cells in a schistosomiasis model (Ding et al. 1993;Flores Villanueva et al. 1994). IL-10 also inhibits iNOS production in macrophages (Cunha et al. 1992) and blocks nuclear factor-κB (NF-κB) activation (Saadane et al. 2005). Tyrosine phosphorylation of the intracellular domains of the IL-10 receptor is known to activate the signaling transducer and activator of transcription-3 (STAT3) pathway (Donnelly et al. 1999). Tarzi et al. (2006) observed that peptide immunotherapy caused induction of IL-10 and suppressor of cytokine signaling 3 (SOCS3) in a bee venom model, and Gao and Ward (2007) has implicated SOCS3 in inflammatory diseases.
Polymorphisms in the human IL10 gene have been associated with the development of asthma (Chatterjee et al. 2005) and lipopolysaccharide (LPS) sensitivity (Schippers et al. 2005). The functional role of pulmonary IL-10 has been studied in experimental animals in response to LPS (Standiford et al. 1995), silica (Huaux et al. 1998), and infectious organisms (Higgins et al. 2003) and as a key mediator in allergy (Akdis et al. 2001) and cystic fibrosis (Saadane et al. 2005). However, the role of IL-10 in O 3 -induced lung inflammation and its underlying mechanism has not been sufficiently studied.
In this study, we tested the hypothesis that targeted deletion of Il10 in mice would enhance O 3 -induced pulmonary inflammation via modulation of expression of inflammatory mediators CD86 and MIP-2 and nuclear transcription factors NF-κB and STAT3. We compared O 3 -induced alterations in pulmonary injury phenotypes and putative downstream molecular events between Il10-sufficient (Il10 +/+ ) and Il10deficient (Il10 -/-) mice. We also used global mRNA expression analyses of lung tissue to identify patterns of gene expression that are modulated by IL-10 after exposure to O 3 . Enhanced O 3 -induced inflammation phenotypes in Il10 -/mice compared with Il10 +/+ mice were consistent with a protective role Background: The mechanisms underlying ozone (O 3 )-induced pulmonary inflammation remain unclear. Interleukin-10 (IL-10) is an anti-inflammatory cytokine that is known to inhibit inflammatory mediators. oBjectives: We investigated the molecular mechanisms underlying interleuken-10 (IL-10)-mediated attenuation of O 3 -induced pulmonary inflammation in mice. Methods: Il10-deficient (Il10 -/-) and wild-type (Il10 +/+ ) mice were exposed to 0.3 ppm O 3 or filtered air for 24, 48, or 72 hr. Immediately after exposure, differential cell counts and total protein (a marker of lung permeability) were assessed from bronchoalveolar lavage fluid (BALF). mRNA and protein levels of cellular mediators were determined from lung homogenates. We also used global mRNA expression analyses of lung tissue with Ingenuity Pathway Analysis to identify patterns of gene expression through which IL-10 modifies O 3 -induced inflammation. results: Mean numbers of BALF polymorphonuclear leukocytes (PMNs) were significantly greater in Il10 -/mice than in Il10 +/+ mice after exposure to O 3 at all time points tested. O 3 -enhanced nuclear NF-κB translocation was elevated in the lungs of Il10 -/compared with Il10 +/+ mice. Gene expression analyses revealed several IL-10-dependent and O 3 -dependent mediators, including macro phage inflammatory protein 2, cathepsin E, and serum amyloid A3. conclusions: Results indicate that IL-10 protects against O 3 -induced pulmonary neutrophilic inflammation and cell proliferation. Moreover, gene expression analyses identified three response pathways and several genetic targets through which IL-10 may modulate the innate and adaptive immune response. These novel mechanisms of protection against the pathogenesis of O 3 -induced pulmonary inflammation may also provide potential therapeutic targets to protect susceptible individuals. for IL-10. Microarray and pathway analyses identified significant differences in inflammatory mediator expression between Il10 +/+ and Il10 -/mice in response to O 3 and suggested novel genetic targets [e.g., cathepsin E (Ctse) and serum amyloid A3 (Saa3)] affiliated with Il10 expression and response to environmental oxidant exposure.

Materials and Methods
Animals. We purchased male Il10 +/+ (C57BL/6) and Il10 -/-(B6.129P2-Il10 tm1Cgn /J) mice (6-8 weeks) from Jackson Laboratories (Bar Harbor, ME). We provided mice with water and pelleted open-formula rodent diet NIH-31 (Zeigler Brothers, Gardners, PA) ad libitum. All experimental procedures were conducted in accordance with approved guidelines from the National Institutes of Health (Institute of Laboratory Animal Resources 1996) and the American Physiological Society (2002). Animals were treated humanely and with regard for alleviation of suffering.
O 3 exposure. We placed mice in individual stainless steel wire cages within a Hazelton 1000 chamber (Lab Products, Maywood, NJ) equipped with a charcoal and high-efficiency particulate air-filtered air supply. We exposed mice to 0.3 ppm O 3 or filtered air for 24, 48, or 72 hr (23.5 hr/day) as described previously (Cho et al. 2007).
Necropsy and BALF analyses. We euthanized mice (intraperitoneal sodium pentobarbital, 104 mg/kg) immediately after exposure. We lavaged the right lung with Hanks' balanced salt solution and processed the lung for cell and total protein (a marker of lung permeability) analyses following Cho et al. (2001). The left lung was snap-frozen in liquid nitrogen for molecular analyses.
Histological analysis. We inflated lavaged right lungs with 10% formalin, removed en bloc, and immersed the lungs in 10% formalin. Details of histological analyses are in the Supplemental Material (doi:10.1289/ ehp.1002182). Immunohistological staining was done using a specific antibody against Ki-67 (ab15580; Abcam, Cambridge, MA); we followed the procedures outlined by Cho et al. (2007).
Real-time quantitative reverse-transcriptase polymerase chain reaction. We isolated total RNA from left lung homogenates using the RNeasy Midi Kit (Qiagen Inc., Valencia, CA) following the manufacturer's instructions and as described by Cho et al. (2007). Details of quantitative reverse-transcriptase polymerase chain reaction (RT-PCR) procedures are outlined in the Supplemental Material (doi:10.1289/ehp.1002182).
Western blot analysis. We prepared total lung protein in RIPA buffer containing protease and phosphatase inhibitors. Details of Western blot procedures and analyses are presented in the Supplemental Material (doi:10.1289/ehp.1002182).
Nuclear protein isolation and electrophoretic mobility shift assay for NF-κB. We prepared nuclear extracts from right lung lobes using a Nuclear Extraction Kit (Active Motif, Carlsbad, CA). Details of the electrophoretic mobility shift assay (EMSA) procedures are in the Supplemental Material (doi:10.1289/ehp.1002182).
Gene array analysis. RNA isolation and Affymetrix GeneChip hybridization. We isolated total RNA from left lung lobes, and the RNA was used after passing quality testing using an Agilent Bioanalyzer 2100 (Agilent Technologies, Inc., Santa Clara, CA, USA). For each treatment group, we performed GeneChip analysis (Affymetrix, Santa Clara, CA) analyses in duplicate (air controls) or triplicate (O 3 exposed). Further details of RNA processing and hybridization are in the Supplemental Material (doi:10.1289/ ehp.1002182).
Data analysis. We normalized and summarized the resulting files (in CEL format) with the Robust Multichip Average method using RMAExpress (http://rmaexpress.bmbolstad.com/). We exported log2 values for analysis using the Spotfire DecisionSite (TIBCO Spotfire, Somerville, MA). Hierarchical clustering identified one outlier, a sample from an Il10 -/mouse exposed to O 3 for 24 hr, which was not included in subsequent analysis. We also applied Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems, Inc., Redwood City, CA), a structured network knowledgebased approach, to evaluate functions and to elect putative interaction and signaling mechanisms through which IL-10 plays a role in pulmonary pathogenesis in response to O 3 . We further analyzed statistically the CEL files by two-way analysis of variance (ANOVA; p < 0.01) using genotype (Il10 +/+ , Il10 -/-) and exposure (air, 24-hr O 3 , 48-hr O 3 , 72-hr O 3 ) as the variables in GeneSpring GX 11.0 Expression Analysis software program (Agilent Technologies). We deposited the raw data   We expressed all data as group means ± SE. We log-transformed PMN data to ensure normal data distribution and equal variance. We assessed differences in effects of O 3 and IL-10 on response phenotypes by two-factor ANOVA. The factors were exposure (air or O 3 ) and genotype (Il10 +/+ or Il10 -/-). Dependent variables were BALF protein concentration, BALF cells, mRNA expression, and protein levels. We used the Student Newman-Keuls post hoc test to compare group means. We performed all statistical analyses using the SigmaStat statistics package (version 3; Systat Software, Inc., Point Richmond, CA). We accepted statistical significance at p < 0.05. Sample sizes are included in the figure captions.  Figure 1A). Total BALF protein concentrations also increased significantly in both genotypes after O 3 , but we found no significant differences between the two genotypes (p = 0.055; Figure 1B).

Lung inflammatory responses. O 3 -induced increases in total numbers of BALF cells were significantly greater in
Histopathological analysis of lung injury. Consistent with BALF phenotypes, we found greater O 3 -induced inflammation in perivascular, peribronchiolar, and terminal bronchial regions in hemotoxylin and eosin (H&E)stained lung tissue sections from Il10 -/mice compared with Il10 +/+ mice ( Figure 1C). Further, the density of Ki67-positive cells at distal perivascular-peribronchiolar areas and centriacinar regions was more prominent in Il10 -/than in Il10 +/+ mice exposed to O 3 , indicating enhanced cellular proliferation ( Figure 1D).
Inflammatory mediator production. We did not detect Il10 mRNA expression in Il10 -/mouse lung homogenates after air or O 3 exposure (data not shown). However, Il10 mRNA was significantly elevated relative to air-exposed controls after 24-hr O 3 (4.4-fold; p < 0.05) but was not significantly different between air-and O 3 -exposed Il10 +/+ mice after 48-and 72-hr exposures (data not shown). Relative to respective air controls, O 3 significantly increased mean BALF concentration of the PMN chemoattractant MIP-2 in both genotypes and was significantly higher in Il10 -/compared with Il10 +/+ mice at 24 hr ( Figure 2A). However, these genotype-dependent differences were not evident at 48-and 72-hr O 3 . TNF-α and iNOS mRNA expression levels were significantly increased in both strains during O 3 , but we found no significant genotypic differences (data not shown).
SOCS3 expression. Because SOCS3 has been proposed as a mechanism of IL-10-mediated protection against inflammatory processes in other models of lung disease (Gao and Ward 2007), we asked whether SOCS3 was differentially expressed in Il10 -/and Il10 +/+ mice. Relative to air controls, Socs3 mRNA expression increased significantly in Il10 +/+ mice after 24-, 48-, and 72-hr O 3 ( Figure 2B); Socs3 expression in Il10 -/mice was increased significantly only after 72-hr O 3 .
CD86 production. Compared with Il10 +/+ mice, we found significantly higher levels of lung CD86 proteins in Il10 -/mice after air and O 3 ( Figure 2C). However, O 3 did not increase CD86 in either genotype, and we found no interaction between genotype and exposure.
Nuclear NF-κB activity and STAT3 activation. To determine whether the protective effect of IL-10 may be mediated in part by differential NF-κB activity, we evaluated nuclear DNA binding activity of total NF-κB and specific p50, a subunit of NF-κB and p65 κB activity in Il10 +/+ and Il10 -/mice. Interestingly, baseline binding activity of total NF-κB and specific p50 κB was slightly greater in Il10 -/mice than in Il10 +/+ mice ( Figure 3A). O 3 effects on activation of total NF-κB and specific p50 κB binding activity were more marked in Il10 -/mice than in Il10 +/+ mice ( Figure 3A). Specific p65 subunit NF-κB activity was significantly increased only in Il10 -/mice after O 3 ( Figure 3A). The ratio of phosphorylated STAT3 (p-STAT3) to STAT3 protein in the lung increased significantly after O 3 compared with air controls in both genotypes, signifying an increase in activated STAT3 protein in response to O 3 ( Figure 3B). However, we found no differences in O 3 -induced increases in the p-STAT3:STAT3 ratio between Il10 -/and Il10 +/+ mice.

Discussion
We investigated the mechanisms of IL-10mediated reductions in O 3 -induced airway neutrophilia in mice. Targeted deletion of Il10 significantly enhanced O 3 -induced pulmonary cellular inflammation, as indicated by significant differences in numbers of infiltrating neutrophils and enhanced cellular proliferation in centriacinar regions compared with Il10 +/+ controls. Nuclear activity of NF-κB and expression of CD86 proteins were also greater in lungs of Il10 -/mice. Furthermore, MIP-2 protein was significantly elevated in response to 24-hr O 3 and was exacerbated in Il10 -/compared with Il10 +/+ mice. Results support a role for IL-10 in protection against O 3 -induced inflammation associated with diminished NF-κB activity and MIP-2 production. Nonbiased (visual, unsupervised) and supervised (ANOVA) gene array analysis identified novel cellular targets that may contribute to the protective mechanism of IL-10 in O 3 -induced inflammation.
O 3 -induced increases in BALF neutrophils and lung cell proliferation were significantly enhanced in Il10 -/mice relative to Il10 +/+ wild-type controls, but the lung permeability response to O 3 was not significantly different between the two genotypes. Interestingly, Reinhart et al. (1999) found that intratracheal instillation of recombinant IL-10 into Sprague-Dawley rats significantly attenuated by 25% protein permeability responses induced by acute O 3 exposure (0.8 ppm, 3 hr) compared with controls. An explanation for the disparate observations is not entirely clear, but the role of IL-10 in the hyperpermeability response may depend on the concentration and duration of exposure to O 3 (0.8 ppm for 3 hr compared with 0.3 ppm for 24, 48, or 72 hr) or may be species specific. Moreover, these results are consistent with previous studies suggesting that inflammatory and per meability responses in this model are regulated through different mechanisms (e.g., Cho et al. 2007).
Upregulation of the PMN chemoattractant MIP-2 has been suggested to regulate initial neutrophilic infiltration in response to O 3 . O 3 -induced increases in MIP-2 were potentiated in the absence of Il10 at 24 hr before the peak PMN influx and therefore may be a mechanism through which IL-10 protects against O 3 -induced inflammation and cell proliferation. Similarly, neutralization Figure 3. (A) EMSA to determine DNA binding activity of total NF-κB (top left) and specific p50 κB (top right) after 0.3 ppm O 3 in Il10 +/+ and Il10 -/mice. Each lane represents nuclear protein pooled from three representative animals of each treatment group and was repeated three times. SB, shifted band; SSB, supershifted band; FP, free probe. Quantified p65 κB determined by ELISA is presented below (means ± SE; n = 3 per group). *p < 0.05, air versus 0.3 ppm O 3 ; **p < 0.05, Il10 +/+ versus Il10 -/-. (B) Phosphorylated STAT3 (p-STAT3) in Il10 +/+ and Il10 -/mice in response to air or 0.3 ppm O 3 . We determined the ratio of p-STAT3 to total STAT3 from whole-lung homogenates by Western blot analysis (means ± SE; n = 3 per group). *p < 0.05, air versus 0.3 ppm O 3 . of IL-10 prior to LPS administration can attenuate production of MIP-2 and associated neutrophilia (Standiford et al. 1995). We observed no significantly different O 3 -induced changes in TNF-α and iNOS between Il10 -/and Il10 +/+ mice, despite genotype-specific differences in NF-κB after O 3 . These results suggest that O 3 -induced increases in TNF-α and iNOS, which are thought to be NF-κB dependent (e.g., Cho et al. 2007), are not modulated by IL-10. These results are inconsistent with those of Lang et al. (2002) and Standiford et al. (1995), who found that TNF-α production is IL-10 dependent in responses to endotoxin. Collectively, these studies may suggest important differences in pulmonary cellular responses to endotoxin and O 3 and warrant further investigation. We also found that lung CD86 protein levels were enhanced in Il10 -/compared with Il10 +/+ mice, independent of O 3 exposure. Soltys et al. (2002) showed that CD86 in alveolar macrophages was significantly enhanced in Il10 -/mice at baseline and suggested that low levels of alveolar macrophage costimulatory molecule expression are maintained at least in part by endogenous IL-10 activity. Koike et al. (2001) found that approximately 8% of alveolar macrophages obtained from rats exposed to 1 ppm O 3 for 3 days were CD86 positive and suggested that small increases in costimulatory activity may be sufficient to regulate O 3 -induced immune responses. Our data similarly indicate that Il10 deficiency permits endogenous CD86 expression, which could explain why we did not detect subtle increases in CD86 in Il10 -/mice after O 3 exposure.
To investigate the potential signaling pathway through which IL-10 protects against O 3 -induced inflammation, we evaluated NF-κB activation in Il10 +/+ and Il10 -/mice. NF-κB is a nuclear protein that regulates transcription of many gene products that activate major pulmonary inflammatory pathways, including responses to O 3 (Cho et al. 2007). IL-10 inhibits NF-κB (Saadane et al. 2005;Spight et al. 2005). We found that O 3 -induced p50 levels and p65 κB binding activity were significantly greater in Il10 -/than in Il10 +/+ mice. The mechanism through which IL-10 inhibits NF-κB may occur indirectly, such as through inhibition of inhibitory kappa kinase, as demonstrated in response to LPS (Saadane et al. 2005;Spight et al. 2005).
STAT3 was associated with IL-10-mediated inhibition of alveolar macrophage activation by LPS (Berlato et al. 2002). Stat3 has also been shown to mediate O 3 -induced inflammation  and pulmonary inflammation and hyperresponsiveness associated with asthmatic responses (Corry 2002). Interestingly, we found no significant genotypespecific differences in the p-STAT3:STAT3 Figure 4. IPA illustrates the three most highly associated networks (A-C) of IL-10-dependent genes in response to O 3 . Red symbols, genes whose profiles were derived from Supplemental Material Figure 5A [see Supplemental Material, Figure 5, Table 2 (doi:10.1289/ehp.1002182)]; yellow symbols, genes with profiles that followed the pattern illustrated in Supplemental Material Figure 5B (see Supplemental Material, Figure 5, Table 2); green symbols, genes with profiles that differ between Il10 +/+ and Il10 -/mice [see Supplemental Material, Figure 5D,  Figure 3B). Similarly, gene array analysis did not detect an interaction among O 3 exposure, Il10 deficiency, and Stat3 gene expression at these time points. These results suggest that O 3 -induced STAT3 activation occurs in an IL-10-independent manner.
To further elucidate the potential mechanism through which IL-10 protects the lung against O 3 -induced inflammation, we used k-means clustering and Ingenuity analyses of microarray expression data. We identified three distinct clusters of molecules, representing three pathways and numerous gene targets. Several genes identified in this analysis (e.g., Il6, Il1ra, Il1) contribute to O 3 -induced pulmonary inflammation in mice (Figure 4).
Other identified genes represent novel targets for O 3 -induced inflammation. For example, both gene expression analyses identified Ctse as differentially expressed between genotypes after exposure to O 3 . Ctse is a an intracellular aspartic protease that is found in immune system cells such as dendritic cells and macrophages and has been implicated in major histocompatibility complex (MHC) class II pathway antigen processing (e.g., Zaidi and Kalbacher 2008). Although Ctse has not been associated with O 3 -induced inflammation, MHC class II molecules have been implicated recently in the PMN response to O 3 exposure in the mouse (Bauer et al. 2009), and it is plausible that CTSE could also be involved in this pathway. SAA and S100A14 have been implicated in inflammatory processes (e.g., Han et al. 2007) or inflammation-related diseases (Chen et al. 2009) and thus provide a rationale for a role in O 3 -induced inflammation.
Our gene array analysis identified Socs3 as being induced in response to O 3 , compared with air controls. RT-PCR analyses confirmed O 3 -induced upregulation of Socs3 and identified a small but statistically significant difference in expression kinetics between Il10 +/+ and Il10 -/mice. Members of the SOCS signaling family, including SOCS3, regulate activation and function of STAT3 (Gao and Ward 2007). Emerging data on activation and function of STAT3 and SOCS3 in the lung during acute inflammation suggest that these molecules may regulate pulmonary inflammation (Gao and Ward 2007), and they have specifically been implicated in allergic airway inflammation (Paul et al. 2009). Given that SOCS3 is a proximal mediator associated with IL-10 ligand binding, the role of SOCS3 in O 3 -induced inflammation necessitates further investigation to determine its specific contribution to the adaptive/innate immune response.
The array analysis implicated a number of novel genes for response to O 3 in Il10 +/+ and Il10 -/mice, including Il33 and Hat1. IL-33 is a member of the IL-1 superfamily and is expressed on epithelial cells. This cytokine has been identified in joints of arthritis patients (Palmer et al. 2009) and is implicated in development and regulation of type 2 T-helper cell-dependent immune responses in sites of mucosal immunity, which may suggest a link between innate and adaptive immune responses after O 3 exposure (Saenz et al. 2008). HAT1, an acetylation enzyme, controls access to DNA transcription by controlling histone protein activation during chromatin assembly (Parthun 2007). Although HAT1 expression has been linked to inflammatory mediators such as interferon-γ and TNF-α (Keslacy et al. 2007), the specific role of HAT1 in response to O 3 remains unclear.

Conclusion
This study supports the hypothesis that IL-10 protects against O 3 -induced lung inflammation and identified several potential mechanisms involved in this response. IL-10 deficiency enhanced O 3 -induced neutrophilic inflammation and injury at the centriacinar region of the lung. Results also suggest that in response to O 3 the effect of IL-10 may be mediated, in part, via modulation of NF-κB, MIP-2, and CD86. Gene array analysis identified three IL-10-mediated expression pathways with genes known previously to be affected by O 3 and novel genes that may contribute significantly to the pathogenesis of O 3 -induced pulmonary inflammation. The present investigation identified novel molecular mechanisms of pulmonary O 3 toxicity, which may provide possible therapeutic targets for attenuating the effects of O 3 in susceptible individuals.