|Home | About | Journals | Submit | Contact Us | Français|
Pulmonary responses to ozone, a common air pollutant, are augmented in obese individuals. Adiponectin, an adipose derived hormone that declines in obesity, has regulatory effects on the immune system. To determine the role of adiponectin in the pulmonary inflammation induced by extended (48–72 h) low dose (0.3 ppm) exposure to ozone, adiponectin deficient (Adipo−/−) and wildtype mice were exposed to ozone or to room air. In wildtype mice, ozone exposure increased total bronchoalveolar lavage (BAL) adiponectin. Ozone induced lung inflammation, including increases in BAL neutrophils, protein (an index of lung injury), IL-6, KC, LIX and G-CSF were augmented in Adipo−/− versus wildtype mice. Ozone also increased IL-17A mRNA expression to a greater extent in Adipo−/− versus wildtype mice. Moreover, compared to control antibody, anti-IL-17A antibody attenuated ozone-induced increases in BAL neutrophils and G-CSF in Adipo−/− but not in wildtype mice, suggesting that IL-17A, by promoting G-CSF release, contributed to augmented neutrophilia in Adipo−/− mice. Flow-cytometric analysis of lung cells revealed that the number of CD45+/F4/80+/IL-17A+ macrophages and γδ T cells expressing IL-17A increased after ozone exposure in wildtype mice, and further increased in Adipo−/− mice. The IL-17+ macrophages were CD11c− (interstitial macrophages), whereas CD11c+ macrophages (alveolar macrophages) did not express IL-17A. Taken together, the data are consistent with the hypothesis that adiponectin protects against neutrophil recruitment induced by extended, low dose ozone exposure by inhibiting the induction and/or recruitment of IL-17A in interstitial macrophages and/or γδ T cells.
Automobile exhaust is a source of many toxic gases and particles, including ozone. Inhalation of ozone has a significant impact on human health and contributes to increased cardiovascular and respiratory mortality (1, 2). In the lung, ozone induces epithelial injury and an inflammatory response that includes a neutrophilic influx and induction of acute phase cytokines including IL(interleukin)-1, IL-6, TNF-α (tumor necrosis factor alpha), as well as the neutrophil chemotactic factors KC (keratinocyte-derived chemokine), MIP-2 (macrophage inflammatory protein), and LIX (LPS induced CXC chemokine) (3–12). Ozone is a trigger for asthma attacks and significantly decreases pulmonary function in asthmatic subjects (13–15). Importantly, responses to ozone are augmented in obese and overweight individuals (16, 17).
Circulating levels of adiponectin, an adipose-derived, energy regulating hormone with anti-inflammatory effects, are reduced in the obese (18–20). Such declines in adiponectin contribute to many obesity-related conditions, including insulin resistance and hypertension (21, 22). Similarly, loss of the anti-inflammatory effects of adiponectin may contribute to obesity-related increases in responses to ozone. For example, macrophages are an important target cell for ozone (4, 23), and TNFα has been shown to be required for the pulmonary neutrophilia caused by ozone (24). Adiponectin decreases LPS-induced TNFα expression in macrophages (25, 26), while augmenting expression of anti-inflammatory molecules such as IL-10 and IL-1RA (27), and skews macrophages from an M1 to an M2 phenotype (28). Consistent with these observations, adiponectin receptors are expressed on most circulating monocytes (29). The ozone-induced influx of neutrophils into the lungs also requires adhesion of neutrophils to endothelial cells (30), and adiponectin has been shown to inhibit TNFα induced expression of VCAM-1, E-selectin, and ICAM-1 on endothelial cells (31). Anti-inflammatory effects of adiponectin have also been demonstrated in the lung in vivo: exogenous administration of adiponectin attenuates allergic inflammation in mice (32), while adiponectin deficiency augments it (33). Nevertheless, pro-inflammatory effects of adiponectin have also been reported (34–36). Adiponectin circulates in multimeric forms: a trimer, a hexamer comprised of two trimers linked by cysteine bonds in the tails, and high-molecular weight (HMW)3 form, composed of 12 to 18 monomers. These isoforms differ in their bioactivity and degradation (19, 37–39). The pro- or anti-inflammatory effects of adiponectin may depend on which adiponectin isoform is present.
If obesity-related declines in adiponectin contribute to the augmented effects of ozone in the obese, we would expect the anti-inflammatory effects of adiponectin to dominate in the setting of ozone exposure. Consequently, adiponectin deficiency would be expected to augment ozone induced inflammation. In contrast, we have reported that adiponectin deficient (Adipo−/−)3 mice exposed to acute ozone (2 ppm for 3 h)3 had decreased neutrophilic inflammation and decreased induction of cytokines and chemokines compared to wildtype mice (40). The acute ozone exposure regimen is frequently used (11, 12, 40–43), because it induces a robust response in mice, which, like other rodents, have a reduced sensitivity to ozone compared to humans (44). However, elevated atmospheric ozone tends to persist not for hours, but for several days or even weeks, the time scale of typical weather patterns (45). Kleeberger et al (7) have developed a “subacute” model where mice are exposed to lower concentrations (0.3 ppm) for longer periods of time (48–72 hours), and many investigators use this as a more realistic model of ozone exposure (3, 6, 7, 11, 46). Importantly, the factors that determine pulmonary responses to short duration, high dose (acute) ozone exposure to ozone are not the same as those that impact responses to longer, lower dose (subacute) ozone (3, 8, 24, 47). For example, TNFα is required for the pulmonary neutrophilia induced by subacute, but not acute ozone exposure in mice (3, 8, 48). Such differences suggest that the impact of adiponectin deficiency on pulmonary responses to subacute versus acute ozone exposure might be different. Therefore, we exposed wildtype and Adipo−/− mice to ozone (0.3 ppm during 24 to 72 h) or to air, and examined pulmonary inflammation and injury. Our data derived using subacute exposure model demonstrated increased ozone-induced inflammation in Adipo−/− mice, indicating that anti-inflammatory activities of adiponectin do indeed dominate under these prolonged exposure conditions.
Interleukin 17A (IL-17A) is involved in neutrophil recruitment to the lung by both infectious and non infectious agents (49–51). In particular, we have reported that IL-17A is induced in the lung by following 3 exposures to ozone (1ppm for 3 h) with 2 day intervals between exposures (51). Importantly, blocking IL-17A either by genetic deletion or via neutralizing antibodies, inhibited the neutrophil influx induced by ozone in this model (51). A source of IL-17A in this situation was invariant NKT cells. To determine whether IL-17A also contributes to the augmented inflammation observed following subacute ozone in Adipo−/− mice, we treated mice with IL-17A neutralizing antibodies.
Adipo−/− mice from Dr. Y. Matsuzawa (Sumitomo Hospital, Osaka, Japan) (52) were bred in house. Gender and age matched C57BL/6J mice (Jackson Laboratories, Bar Harbor, USA) were used as wildtype controls. All protocols were approved by The Harvard Medical Area Standing Committee on Animals. Mice were 11–13 weeks old at the time of exposure.
Mice were housed in microisolator cages, with their HEPA filter removed, and given with water and food ad libitum. The cages were then placed in a 145 l stainless steel and Plexiglas exposure chamber. Mice were exposed for up to 72 hours to 0.3ppm of ozone as previously described (6, 7). Ozone was generated from medical grade oxygen by an in house electric ozone generator. Ozone levels were continuously monitored as described before (40). Air exposed mice were exposed concurrently, in a separate, but identical chamber.
In the first cohort of experiments, Adipo−/− and wildtype mice were exposed to air or 0.3ppm of ozone for up to 72 hours (subacute ozone). At the end of exposure, mice were euthanized with an overdose of pentobarbital. The chest cavity was opened, venous blood was drawn by a cardiac puncture of the right ventricle, and serum was prepared. The trachea was then cannulated and bronchoalveolar lavage (BAL) was performed using two instillations of 1ml PBS, pH 7.4. The left lung lobe was harvested and immersed in RNAlater solution (Qiagen, Valencia, CA) for 24 hours and then transferred to a −20°C freezer until isolation of total RNA was performed. In the second set of experiments, Adipo−/− and wildtype mice were treated with an i.p. injection of 100µg in 100 µL of either anti IL-17 neutralizing antibody (Rat IgG2A, clone 50104, MAB421, R&D Systems, Minneapolis, MN) or isotype control antibody (clone 54447, MAB006, R&D Systems, Minneapolis, MN) resuspended in sterile saline. After thirty minutes, mice were exposed to 0.3ppm of ozone for 48 hours, and then blood, BAL fluid, and lungs were harvested as described above. In the third series of experiments, Adipo−/− and wildtype mice were exposed to ozone or air for 72 hours. The animals were then euthanized and the lungs harvested and cells isolated for flow cytometry to identify cellular sources of IL-17A.
BAL fluid was spun at 1500 rpm at 4°C. The cell pellet was resuspended with 1mL of PBS and total cells were counted using a hemocytometer. Cytospin preparations were obtained, stained with hematoxylin and eosin and used to obtain differential cell counts on at least 300 cells. BAL fluid was aliquoted and stored at −80°C until assayed for total protein (a measure of ozone induced lung epithelial injury), adiponectin, and inflammatory cytokines and chemokines.
Total adiponectin, MIP-2, KC, IL-1Ra, IL-6, TNF-α, granulocyte colony stimulating factor (G-CSF), and LIX levels in the BAL fluid or serum were measured by commercially available enzyme-linked immunosorbent assay kits (R&D systems and BD-Bioscience, San Jose, CA). BAL protein was measured by Bradford assay (BioRad, Hercules, CA).
BAL and serum adiponectin isoforms were resolved by a non-reducing western-blot method as previously described (40). Briefly, 30μL of BALF in SDS sample buffer was separated on 1.5mm 4–12% PAGE in MES buffer (Invitrogen, Carlsbad, CA), transferred to a nitrocellulose membrane, and probed with anti-adiponectin antibody. Adiponectin isoforms were identified via their position relative to molecular weight standards.
The left lung was homogenized and total RNA purified using a RNeasy Column kit (Qiagen) and treated with DNase II to remove genomic DNA. RNA amount and purity was determined at 260 and 280nm by using a small volume spectrophotometer (Nanodrop, Thermo Scientific, Waltham, MA, USA). 1 µg of RNA was converted into cDNA using “Super Script III First-strand amplification kit for qRT-PCR” (Invitrogen, Carlsbad, USA). IL-17A, IL10 and IkappaB zeta (IκBζ) mRNA were quantitated by real time PCR (7300 Real-Time PCR Systems, Applied Biosystems, Carlsbad, CA, US) using intron-spanning primers (IL-17A sense: 5’ccagggagagcttcatctgt3’ and antisense:5’aggaagtccttggcctcagt3’; IL-10 sense: 5’agccgggaagacaataactg3’ and antisense: 5’gttgtccagctggtcctttg3’; IκBζ sense (53): 5’cagttgcctgtctttcgtga3’ and antisense: 5’tccaactgtgtcacccgata 3’) and SYBR-green detection. IL-10, IκBζ, or IL-17A were normalized to 18S expression using the ΔΔCt method (54).
After exposing mice to 0.3ppm ozone or air as described above, BAL was performed, and the lungs were flushed by instilling 10mL of ice cold PBS into the right ventricle after sectioning the wall of the left ventricle. The lungs were then harvested and perfused with PBS containing 2% FCS. Lung tissue was minced and incubated with collagenase type 4 (Worthington Biochemical Corp, 0.6 mg/mL) and Golgi stop (BD bioscience) in RPMI at 37°C. Harvested cells were counted, and preincubated with anti-Fcγ blocking mAb (2.4G2) and washed. Cells were then stained with anti–mouse PE-Texas red conjugated CD45 (clone 30-F11; Invitrogen) to identifiy hematopoetic cells. To examine lymphocytes, cells were stained with PerCPCy5.5-conjugated CD3 (clone 145-2C11; BD bioscience), CD4 T cells were identified with Alexa Fluor 750 – conjugated CD4 mAb (clone RM4-5; eBioscience), and γδ T cells were identified with anti-mouse PE-conjugated TCRγδ mAb (clone GL3, eBioscience). To examine macrophages, cells were stained with FITC-conjugated F4/80 mAb (clone BM8; Biolegend), Alexa Fluor 700-conjugated CD11b mAb (clone N418; eBioscience) and Alexa Fluor 750-conjugated CD11c mAb (clone M1/70; Biolegend). For intracellular staining for IL-17A, cells were permeabilized (Cytofix/Cytoperm, BD Biosciences), and incubated with Alexa Fluor 647-conjugated IL-17 antibody or isotype control (Alexa Fluor 647-conjugated rat IgG2a (eBioscience). To confirm that the IL-17A observed in macrophages was localized intracellularly rather than on the cell surface, control experiments were conducted in which we did not perform the permeabilization step. To examine IL-17A expression in Th17 and γδ T cells, aliquots of cells were incubated either with or without PMA (100ng/mL) and ionomycin (500ng/mL) for 5 hours prior to staining for flow cytometry. All labeled cells were passed through a BD Canto flow cytometry (BD Bioscience) and data analyzed with FlowJo software (Tree Star, Inc.).
Data was analyzed by factorial ANOVA with Fisher-LSD as post-hoc (STATISTICA, StatSoft, Tulsa, USA). BAL cells were log transformed in order to normalize the data. A p value <0.05 was considered statistically significant.
In wildtype mice, exposure to ozone (0.3ppm for 72 hours) had no effect on serum adiponectin (fig.1A). In contrast, ozone increased BAL adiponectin (fig.1B), likely as a result of increased permeability of the lung epithelial barrier (see below). The increase in BAL adiponectin was time dependent, being absent 12 h after initiation of exposure and increasing to maximum at about 48 hours after the onset of exposure (fig.1C). Ozone exposure for 12, 24 or 48 h had no effect on serum adiponectin similar to the results obtained at 72 h (data not shown). There were also changes in the adiponectin isoform distribution in the BAL fluid. In air exposed mice, HMW adiponectin dominated BAL fluid, with slightly lesser amounts of hexameric adiponectin, and no detectable trimeric isoform. Ozone exposure for 72 h caused marked increases in both hexameric and trimeric adiponectin, with only comparably small increases in HMW adiponectin (fig.1D).
Ozone caused a significant increase in the number of neutrophils and macrophages in BAL fluid (fig.2). In wildtype mice, neutrophils were increased as early as 24 hours after initiation of exposure, peaking at 48 hours (fig.2A). BAL neutrophils were increased in Adipo−/− versus wildtype mice at all 3 time points examined. In wildtype mice, BAL macrophages increased by 48 hours of exposure and remained elevated thereafter (fig.2B). In Adipo−/− mice, 72 hours of ozone exposure were required for macrophages to increase, ultimately reaching the same level as in wildtype mice. Ozone causes damage to the pulmonary epithelial barrier resulting in increased permeability and an increase in total BAL protein as serum proteins leak into the airspaces (55). In wildtype mice, BAL protein increased by 24 h of exposure, peaked by 48 hours and remained elevated thereafter (fig.2C). BAL protein was greater in Adipo−/− versus wildtype mice.
BAL cytokines and chemokines implicated in neutrophil recruitment were measured in mice exposed to air or to ozone for 72 hours (fig 3). In wildtype mice, BAL IL-6, TNF-α, and KC were significantly increased in ozone versus air exposed mice (p<0.05), whereas MIP-2, LIX, and G-CSF were not. In Adipo−/− mice, ozone caused significant increases in BAL LIX, and G-CSF, as well as IL-6, TNF-α, and KC. Except for BAL TNF-α, each of these moieties was significantly greater in ozone exposed Adipo−/− versus wildtype mice (p<0.05).
Because others have reported that adiponectin can augment expression of the anti-inflammatory cytokine, IL-10 (27), and because Backus et al (56) have reported increased pulmonary inflammation in IL-10 deficient mice exposed to the same regimen of ozone exposure as we used, we also measured lung IL-10 mRNA expression. We did not observe any significant effect of ozone exposure on IL-10 mRNA (data not shown), and there was no effect of genotype, suggesting that the increased inflammation observed in Adipo−/− mice was not result of loss of adiponectin-induced IL-10 expression.
We measured the endogenous IL-1 receptor antagonist (IL-1Ra) since it is known to be induced by subacute ozone (57), and it antagonizes the effects of IL-1(58), which is required for the neutrophil influx induced by subacute ozone (47). Further, IL-1Ra expression in macrophages is induced by adiponectin (27). We reasoned, therefore, that ozone-induced expression of IL-1Ra might be impaired in Adipo−/− mice, leading to greater IL-1 signaling, which may account for the increase in neutrophils (fig.2A). Our results do not support this hypothesis, since IL-1Ra was actually elevated to a greater degree in Adipo−/− versus wildtype mice exposed to ozone (fig.3G).
Compared to air, ozone increased IL-17A mRNA expression in Adipo−/− mice (fig.4A). A similar trend was observed in wildtype mice but did not reach statistical significance. Moreover, after 48 h of ozone exposure, IL-17A mRNA expression was significantly greater in Adipo−/− than in wildtype mice.
To determine whether the greater induction of IL-17A in the Adipo−/− mice contributed to their enhanced neutrophil recruitment, both Adipo−/− and wildtype mice were treated with anti-IL-17A neutralizing antibody or isotype IgG control antibody and exposed to ozone for 48 hours. After control antibody treatment, BAL neutrophils were higher in Adipo−/− than in wildtype mice, consistent with our observations in Fig.2. In Adipo−/− mice, BAL neutrophils were significantly reduced after anti-IL-17A antibody compared to isotype control antibody (fig.4B). However, neutralization of IL-17A did not significantly affect neutrophil recruitment in wildtype mice. In contrast to its effects on neutrophil recruitment, anti-IL-17A antibody treatment did not affect ozone induced increases in BAL macrophages (fig.4D) or protein (fig.4C), IL-6, or LIX in either genotype studied (fig.4E and G). BAL KC was paradoxically increased in Adipo−/− treated with anti-IL-17A antibody (fig.4F), whereas BAL G-CSF was significantly reduced in ozone exposed Adipo−/− mice treated with anti-IL-17A antibody (fig.4H).
Flow cytometry was performed to determine which cell types in the lung were the source of the IL-17A that was induced following ozone exposure. Ozone exposure (0.3 ppm for 72 h) increased the number of IL-17A expressing CD45+ cells in the lung (fig.5A) consistent with the increase in IL-17A mRNA (fig.4A). Moreover, the number of IL-17A expressing cells was significantly higher in Adipo−/− than in wildtype mice (fig.5A).
Staining for F4/80 antigen, a macrophage marker, indicated the presence of IL-17A positive macrophages in the lung tissue (fig.5B). In contrast, BALmacrophages from ozone exposed mice were not IL-17A positive (see supplemental figure 1B). BAL macrophages from ozone exposed mice were almost exclusively F4/80+CD11c+, whereas the lung tissue had both F4/80+CD11c+ “alveolar” macrophages and F4/80+CD11c− “interstitial” macrophages and it was the latter that expressed IL-17A (fig. 5B). The percentage of CD45+/F4/80+/IL-17A+ cells in air exposed WT and Adipo−/− was significantly different (1.0 ± 0.1 and 2.1 ± 0.1 %, respectively, p<0.05). Ozone induced a significant increase in IL-17A+ macrophages in both WT and Adipo−/− mice to 2.5 ± 0.3 % and 3.2 ± 0.3 %, respectively (p<0.05 in each case). There was a significant increase in the total number of CD45+/F4/80+/IL-17A+ cells in ozone versus air exposed lungs, and significantly higher numbers of CD45+/F4/80+/IL-17A+ in Adipo−/− versus WT mice exposed to ozone (fig.6A). In non-permeabilized cells, no CD45+/F4/80+/IL-17A+cells were observed, confirming the intracellular source of this IL-17A (supplemental figure 2). The observation that surface labeling for IL17R1 was observed on CD11c+ but not CD11c− macrophages also suggests that the IL-17 expression observed in CD45+/F4/80+/CD11c− cells was intracellular and not surface associated(data not shown).
To examine the impact of ozone and adiponectin deficiency on IL-17A expression in T lymphocytes, cells were incubated with PMA and ionomycin for 5 hours before IL-17A staining and permeabilization. Virtually no IL-17A was observed in either CD4 T cells (CD45+/CD3+/CD4+) or γδ T cells (CD45+/CD3+/TCRδ+) in the absence of stimulation (supplemental figure 1A and figure 5C respectively). IL-17A was observed in some CD45+/CD3+/CD4+ cells after stimulation with PMA and ionomycin. However, these Th17 cells accounted for only a small percentage of CD4+ cells even after ozone exposure (supplemental figure 1A). In contrast, a much larger percentage of γδ T cells expressed IL-17 after stimulation with PMA and ionomycin (fig.5C). Ozone increased total number of γδ T cells (fig.6B) in lungs from WT and to a greater extent Adipo−/− mice. IL-17A+ γδ T cells were also significantly greater in ozone exposed Adipo−/− versus WT mice (fig.6C). However, ozone had no effect on the number of IL-17A positive iNKT cells or neutrophils (data not shown).
IκBζ or MAIL (molecule possessing ankryn-repeats induced by LPS) is a transcription factor that is required for the induction of IL-17A in CD4+ T cells (59) and is also required for mediating IL-17A induced expression of certain genes (60–62). Factorial ANOVA indicated an increase in IκBζ mRNA expression in Adipo−/− versus wildtype mice (Fig 7, p=0.004) regardless of exposure.
Our data indicate that adiponectin is present in the BAL fluid and is markedly increased following subacute ozone exposure (fig.1B,C). Furthermore, adiponectin is functionally important, since ozone-induced inflammation and injury were augmented in mice deficient in adiponectin (fig.2 and and3).3). Importantly, ozone increased the number of IL-17A expressing CD11c- macrophages in the lung tissue (fig.5B and and6A).6A). Greater numbers of IL-17A+ macrophages occurred in Adipo−/− versus wildtype mice. Ozone also caused a greater increase in the number of PMA and ionomycin stimulated IL-17+ γδ T cells in lungs of Adipo−/− versus wildtype mice (fig.6C). Importantly, neutralization of IL-17A reduced ozone-induced neutrophilic inflammation, but not injury, in Adipo−/− mice (Fig.4). Taken together, the data are consistent with the hypothesis that during extended ozone exposure, adiponectin exerts anti-inflammatory effects that inhibit the expression of IL-17A in the lung, thus limiting the influx of neutrophils. The data also suggest that interstitial macrophages and/or γδ T cells are the likely source(s) of this IL-17.
We observed a marked increase in the amount of adiponectin present in the lung following ozone exposure, as well as differences in the adiponectin isoform distribution (fig.1). Adiponectin is expressed almost exclusively in adipocytes, and has to enter the lung from the circulation. The increase in BAL adiponectin was not the result of increased serum adiponectin, which did not change with ozone (fig.1A). Ozone impairs both epithelial and endothelial integrity (63) and it is likely that both the increased levels of BAL adiponectin (fig.1B and C) and the altered adiponectin isoform distribution (fig.1D) are the result of the marked increase in lung permeability that occurs following subacute ozone exposure (fig.2B). Given the very high molecular weight, especially of the hexameric and HMW adiponectin isoforms, movement of adiponectin from the blood into a normal air exposed lung is unlikely to be via simple diffusion through gaps between endothelial cells: such a diffusive process would result in the greatest influx of trimeric adiponectin, since it is smallest and hence, should diffuse most easily. Instead, HMW adiponectin was the isoform most plentiful in BAL fluid, with slightly lesser amounts of hexameric adiponectin and virtually undetectable amounts of the trimeric isoform (fig.1D), consistent with a previous report from our lab (40). Indeed, in the normal lung, the adiponectin binding protein, T-cadherin, appears to be required for adiponectin transit into the lung (40). In contrast, under conditions of the marked increase in permeability which characterizes the ozone exposed lung, adiponectin may be able to diffuse. The observation that it is the smallest isoform, trimeric adiponectin, whose concentration undergoes the most marked increase in BAL fluid following subacute ozone exposure (fig.1D), suggests that this is indeed the case.
These ozone-induced changes in adiponectin concentration and isoform distribution appear to be functionally important, since exposing mice deficient in adiponectin to ozone resulted in a significant increase both in BAL neutrophils (fig.2), and in IL-6, KC, LIX, G-CSF, and IL-17A (figs.3 and and4),4), chemotactic factors that can contribute to neutrophil recruitment. The data suggest that under conditions of subacute ozone exposure, adiponectin has anti-inflammatory effects, decreasing the induction of pro-inflammatory mediators and subsequent neutrophilic inflammation. In contrast, our previous data indicate that following an acute 3 hour exposure to a much higher concentration of ozone (2ppm), adiponectin deficiency does not augment, but rather attenuates both the neutrophil recruitment and the cytokine and chemokine induction that occur (40). Consistent with these observations, adiponectin deficiency causes a reduction in IL-17A mRNA expression induced by this acute ozone exposure (data not shown). Other differences in the factors that modify responses to acute versus subacute ozone have been reported. For example, Kleeberger et al. have demonstrated, using mice backcrossed from ozone susceptible and resistant strains, that the genetic factors that control neutrophil influx induced by subacute versus acute ozone are not the same (7, 24, 48). Further, deficiency in IL-1R1, the main signaling receptor for both IL-1β and IL-1α, or in TNF-α signaling, causes a marked reduction in the neutrophil influx that occurs with subacute ozone, but does not alter responses to acute ozone (3, 8, 24, 47). The mechanistic basis for differences in the role of adiponectin under conditions of acute versus subacute ozone is not clear. However, both pro- and anti-inflammatory effects of adiponectin have been reported by others (see Fantuzzi et al (64) for review). It is possible that differences in the amount or isoform distribution of adiponectin in the lung under the two ozone exposure conditions may play a role. We did not observe any increase in BAL adiponectin in wildtype mice following acute ozone exposure (40), whereas subacute ozone caused profound increases in BAL adiponectin (fig.1). We have not measured the adiponectin isoform distribution in BAL of mice exposed to acute ozone, but the fact that the increase in lung permeability was much more limited in those mice (about a 2-fold increase in BAL protein (40)) versus mice exposed subacutely (about 5 fold, fig.2C) suggests that the ratio of the trimeric to HMW isoforms following the two different ozone exposures is likely to be quite different. The HMW isoform of adiponectin has been shown to cause NF-κB activation, whereas in the same cell type, trimeric adiponectin does not (65, 66).
IL-6, KC, LIX, G-CSF, and IL-17A were each induced by ozone exposure to a greater extent in Adipo−/− versus wildtype mice (figs 3 and and4).4). Each of these cytokines and chemokines has been shown to play a role in neutrophil recruitment by various factors, in some cases including ozone exposure (6, 41, 50, 51, 67), albeit with different ozone exposure regimens. We chose to focus on IL-17A because it is known to play a role in neutrophil recruitment in response to diverse stimuli (infection, treatment with LPS, allergen challenge) (50, 67–74). In addition, previous data from our group indicates that IL-17A is required for the neutrophil recruitment that occurs after three repeated exposures to 1ppm ozone (each for 3 h) over the course of 5 days (51). Indeed, our data indicate that inhibiting IL-17A with neutralizing antibodies attenuates the neutrophil influx that occurs with ozone exposure in Adipo−/− mice and normalizes differences in neutrophil recruitment between Adipo−/− and wildtype mice (fig.4). The neutrophil recruitment induced by IL-17A is not direct but is typically mediated by induction of other factors that recruit neutrophils such as IL-6, KC, LIX, and G-CSF (6, 41, 67). Our results indicate that IL-17A induction of G-CSF is likely to account for the increased neutrophil recruitment induced by adiponectin deficiency, since anti-IL-17A antibodies inhibited G-CSF, but not LIX, IL-6, or KC expression (fig.4).
We were somewhat surprised to find that CD11c− macrophages and γδ T cells were the sources of the IL-17A in the lung after subacute ozone exposure (fig.5). Previous data with ozone exposed mice had indicated that iNKT cells were involved (51), whereas in the current study exposure to 0.3ppm ozone for 72 hours did not induce recruitment of iNKT cells to lung. The primary difference between the studies is the ozone exposure regimen: ozone exposure in this study was to lower concentrations (0.3ppm) versus 1ppm in the study by Pichavant et al (51), although it is also possible that the continuous nature of the 72 h exposure used here results in adaptive mechanisms that do not occur with the briefer intermittent (3 h×3) exposures used in the Pichavant study (51). IL-17A expression in macrophages is not without precedent. IL-10 deficient macrophages stimulated with LPS produce IL-17A (75), and exposure to chitin can also induce IL-17A in macrophages (76). IL-17A positive macrophages are also observed in the lungs of mice with allergic airways inflammation (77) and after RSV3 infection (78).
The IL-17A+ macrophages that were elicited in the lung following subacute ozone exposure were CD11c negative. Indeed, IL-17A+ macrophages were observed in the lung tissue but not in BAL fluid, which contains mostly CD11c+, alveolar macrophages. CD11c− macrophages are considered “inflammatory” or “interstitial” macrophages (79). The observation that these IL-17A+ macrophages were increased to a greater extent in ozone exposed Adipo−/− versus wildtype mice suggests either that adiponectin inhibits the expression of IL-17A in these macrophages, or that in the lung, adiponectin reduces the recruitment or proliferation of IL-17A expressing macrophages.
Compared to wildtype, Adipo−/− mice had increases in both total and IL-17+ γδ T-cells even after air exposure, and subacute ozone significantly increased IL-17+ γδ T-cells in Adipo−/− mice (fig.6A). Production of IL-17A by γδ T-cells has also been shown to occur in allergic inflammation and infection (80–82). Although the precise stimulus that induces the increased numbers of IL-17+ γδ T cells in the lungs following ozone exposure (Fig 6A) is not established, it is important to note that IL-17+ γδ T cells do express TLR2 (83). Others have reported that TLR2 is required for other aspects of the pulmonary response to ozone (84). In adipose tissue, γδ T cells are the primary IL-17 producing T cell subset (85), consistent with a role for adipokines in regulating IL-17 in these cells. However, the definitive mechanism whereby adiponectin deficiency affects IL-17+ γδ T cells remains to be established. Finally, it is important to note that the number of IL-17+ γδ T-cells, even after stimulation, was lower than the number of IL-17+ interstitial macrophages (compare figures 6A and 6B).
The observation that IκBζ was increased in Adipo−/− versus wildtype mice (fig.7) suggests that it is the expression of IL-17A that may be affected by adiponectin. IκBζ is a transcription factor that acts synergistically with RORγt to induce IL-17A in CD4+ T-cells (59). The observation that IL-17A can also induce IκBζ expression (62), suggests that a positive feedback loop between IL-17A and IκBζ may also exist in some cells. We observed increased expression of IκBζ mRNA in the lungs of Adipo−/− versus wildtype mice regardless of exposure (fig.7). Since the ability of IL-17 to induce expression of other genes, including IL-23R, IL-21, β-defensin 2, lipocalin, and IκBζ itself, also requires IκBζ (59–62), adiponectin may also inhibit the downstream effects of IL-17A in ozone exposed mice.
In conclusion, adiponectin plays an important role in the pulmonary inflammatory response to subacute ozone exposure. Adiponectin, exerts anti-inflammatory effects leading to decreased neutrophil recruitment and decreased expression of pro-inflammatory cytokines and chemokines. There was a marked increase in IL-17 mRNA expression in lungs of Adipo−/− mice following subacute ozone exposure. This IL-17 was required for the augmented ozone-induced neutrophil influx that occurred in Adipo−/− mice. Interstitial macrophages and/or γδ T cells were the main source of this cytokine. IL-17 was also required for ozone induced G-CSF release into BAL fluid suggesting that IL-17 induced increases in G-CSF may participate in the augmented recruitment and/or survival of neutrophils in Adipo−/− mice.
This study was supported by the U.S. National Institute of Health grants HL-084044, ES-013307, HD25938, HL-077499, and ES-00002. CH was supported by a Career Development Fellowship from Children's Hospital Boston.
3Abbreviations used in this manuscript: Adipo−/−, adiponectin deficient mice; BAL, bronchoalveolar lavage; HMW, high molecular weight adiponectin; ppm, parts per million; RSV, respiratory syncytial virus; PMA, phorbol myristate acetate.