Regulation of IL-17A expression in mice following subacute ozone exposure

Abstract

Exposure to subacute ozone (O₃) causes pulmonary neutrophil recruitment. In mice, this recruitment requires IL-17A. Ozone also causes expression of IL-23 and IL-1, which can induce IL-17A. The purpose of this study was to examine the hypothesis that IL-23 and IL-1 contribute to IL-17A expression and subsequent neutrophil recruitment after subacute O₃ exposure. Wild-type, IL-23^−/−, and Flt3l^−/− mice were exposed to air or 0.3 ppm O₃ for 72 h. Flt3l^−/− mice lack conventional dendritic cells (cDC) that can express IL-23 and IL-1. Other wild-type mice were pre-treated with saline or the IL-1R1 antagonist anakinra prior to O₃ exposure. After exposure, bronchoalveolar lavage (BAL) was performed and lung tissue harvested. The results indicated that pulmonary Il17a mRNA abundance and IL-17A⁺ F4/80⁺ cells were significantly reduced in O₃-exposed IL-23^−/− vs in wild-type mice. In contrast, anakinra had no effect on Il23a or Il17a pulmonary mRNA abundance or on BAL concentrations of the neutrophil survival factor G-CSF, but anakinra did reduce BAL neutrophil numbers, likely because anakinra also reduced BAL IL-6. Compared to air, O₃ caused a significant increase in DC numbers in wild-type, but not in Flt3^−/− mice. However, there was no significant difference in Il23a or Il17a mRNA abundance or in BAL neutrophil count in O₃-exposed Flt3^−/− vs in wild-type mice. From these results, it was concluded that IL-23 but not IL-1 contributes to the IL-17A expression induced by subacute O₃ exposure. Induction of IL-23 by O₃ does not appear to require cDC.

Keywords:

• IL-23
• IL-1
• dendritic cells
• neutrophils
• IL-6

Introduction

Exposure to the air pollutant, ozone (O₃), causes asthma symptoms (Gent et al. 2003), increases susceptibility to respiratory infections (Stanek et al. 2011) and even increases the risk of mortality (Bell et al. 2004; Ito et al. 2005). These responses to O₃ may be consequences of lung epithelial barrier disruption (Kehrl et al. 1987; Lang et al. 2008), impaired immune cell function within the respiratory tract (Gilmour et al. 1991, 1993) and recruitment of neutrophils (PMN) to the lungs (Alexis et al. 2013; Barreno et al. 2013; Kasahara et al. 2013).

This laboratory has previously reported that, in mice, IL-17A contributes to the PMN recruitment observed following subacute O₃ exposure (0.3 ppm for 24–72 h), at least in part via effects of IL-17A on expression of G-CSF, a PMN survival factor (Kasahara et al. 2012; Mathews, Williams, et al. 2014). Sources of IL-17A in the lung after subacute O₃ include macrophages and γδ T-cells (Kasahara et al. 2012; Mathews, Williams, et al. 2014). IL-23 can promote IL-17A expression in T-cells including γδ T-cells (Gaffen et al. 2014). The receptor for IL-23 is widely expressed on immune cells (Parham et al. 2002) and binding of IL-23 to its receptor induces the activation of multiple transcriptional pathways that lead to the expression of IL-17A (Harrington et al. 2006). IL-23 expression is induced in the lungs following subacute O₃ exposure (Mathews, Williams, et al. 2014), but its role in subsequent IL-17A expression and PMN recruitment has not been established. IL-1 also has the capacity to increase IL-17A expression in the lungs (Chung et al. 2009; Gasse et al. 2011; Kim et al. 2014), perhaps because IL-1 can induce IL-23 expression (Harris et al. 2008; Shainheit et al. 2008). Both IL-1α and IL-1β are induced in the lungs following subacute O₃ exposure and IL-1 contributes to the subsequent PMN recruitment induced by O₃ (Johnston et al. 2007), but whether IL-1-induced changes in IL-17A expression, either directly or via changes in IL-23, contribute to these effects on PMN recruitment has not been established.

The present study was done to validate the hypothesis that IL-23 and IL-1 contribute to the production of IL-17A after subacute O₃ exposure. To address this hypothesis, IL-23^−/− and wildtype (WT) control counterparts were exposed to O₃ (0.3 ppm O₃ for 72 h) and ensuing pulmonary inflammation was measured. WT mice were also treated with the IL-1R antagonist anakinra or saline prior to O₃ exposure to determine whether effects of IL-1 were mediated via an ability to induce IL-23 and/or IL-17A expression.

Dendritic cells (DC) are among the sources of IL-23 in the lungs (Aggarwal et al. 2003; Zhang et al. 2015). While the role of DC as antigen-presenting cells is perhaps more widely appreciated, DC also participate in innate immune responses. DC express toll-like receptors, nod-like receptors and the CD1 family of receptors, enabling DC to quickly respond to pathogen- and damage-associated molecular patterns (DAMPs). DAMPs such as hyaluronan, nucleosides and the high-mobility box group 1 protein are produced in the lungs following O₃ exposure (Hollingsworth et al. 2007) and the number of DC in the lungs increases after O₃ exposure (Brand et al. 2012). Whether DC contribute to the IL-23 and IL-1 induced following O₃ has not been established. However, the observations that DC lie directly adjacent to γδ T-cells in the lungs (Wands et al. 2005) and that γδ T-cells are important sources of IL-17A required for O₃-induced PMN recruitment (Mathews, Williams, et al. 2014) suggest DC may indeed be involved.

Consequently, the present study also examined the hypothesis that DC are an important source of the IL-23 driving IL-17A expression after subacute O₃ exposure. To address this hypothesis, fms-like tyrosine kinase-3 ligand null (Flt3l^−/−) mice and their WT controls were exposed to O₃ (0.3 ppm O₃ for 72 h) or air and ensuing pulmonary inflammation was measured. Flt3l is required for differentiation of the common myeloid progenitor of cDC (Waskow et al. 2008). Thus, Flt3l^−/− mice lack conventional DC (cDC) that represent the majority of lung DC (Hammad and Lambrecht 2011).

Materials and methods

Animals

IL-23^−/− mice on a C57BL/6 background (Ghilardi et al. 2004) were kindly provided by Nico Ghilardi (Genentech, San Francisco, CA). IL-23^−/− and littermate control wild-type (WT) mice were bred from IL-23^+/− breeding pairs. IL-23^−/− mice display no overt phenotype. Development, fertility, size and weight are all normal and in the absence of stimulation there is no apparent effect on immune system development aside from a small increase in CD8⁺ DC (Ghilardi et al. 2004). Flt3l^−/− mice (McKenna et al. 2000) and gender- and age-matched WT (C57BL/6) control mice were purchased from Taconic Biosciences (Hudson, NY). For other experiments, including experiments involving anakinra treatment, WT (C57BL/6) mice were purchased from The Jackson Laboratory (Jackson Labs, Bar Harbor, ME). All mice were provided ad libitum access to a standard mouse chow diet and filtered tap water. All mice were 8–12-weeks-old at the time of study and were acclimated for 1 week before exposure. All mice were housed in a vivarium maintained at 21–22 °C with a 48–50% relative humidity and a 12-h light:dark cycle. The Harvard Medical Area Standing Committee on Animals approved all aspects of this study.

Protocol

To study the role of IL-23 in responses to O₃, WT and IL-23^−/− mice were exposed to 0.3 ppm O₃ for 72 h, as previously described (Kasahara et al. 2012). Following exposure, mice were immediately euthanized by overdose with an intraperitoneal injection of sodium pentobarbital. The trachea was then exposed, canulated and bronchoalveolar lavage (BAL) was performed in situ. The lungs were then cleared of blood by perfusing the right ventricle with 10 ml of ice-cold phosphate-buffered saline (PBS, pH 7.4) after creating an incision in the left ventricle. The right lung was used for flow cytometry and the left lung lobes were placed in RNAlater (Qiagen, Germantown, MD) for subsequent preparation of RNA for real time PCR.

To study the role of IL-1α and IL-1β, a different set of WT mice were treated with a recombinant form of human IL-1 receptor antagonist anakinra or with saline. Anakinra is efficacious in mice (Alexander et al. 1991) and both IL-1α and IL-1β require IL-1R1 for signaling (So et al. 2007). Anakinra (100 mg/kg, intraperitoneally, Swedish Orphan Biovitrium AB, Stockholm, Sweden) or an equal volume of saline was administered 24 h before, immediately prior to and every 24 h during the O₃ exposure protocol for a total of four injections. The dose of anakinra used here was based on prior studies such as Alexander et al. (1991). These animals were euthanized immediately after O₃ exposure and BAL and lung tissue collection performed as above. A similar protocol to that used for IL-23^−/− mice and their controls was used to examine the role of cDC, using Flt3l^−/− and their WT controls.

Bronchoalveolar (BAL) lavage

BAL was performed on each mouse by instilling two aliquots of 1 ml ice-cold PBS into the lung via the trachea. After centrifugation, BAL supernatant was stored at −80 °C until assayed. BAL cells were centrifuged onto glass slides and stained using the Hema3 Stain kit (equivalent to Wright-Geimsa). At least 300 cells were counted. BAL cytokines and chemokines were measured using a multiplex assay (Eve Technologies, Calgary, Alberta, Canada). Total BAL protein was measured using a Bradford assay (Bio-Rad, Hercules, CA).

Flow cytometry

The left lung of each mouse was harvested and placed on ice in RPMI 1640 media containing 5% fetal bovine serum (FBS) and 20 mM HEPES (Life Technologies, Grand Island, NY). The lungs were then minced, returned to the media and incubated at 37 °C with 2.5 mg/ml collagenase IV (Roche, Mannheim, Germany) for 30 min while rotating. Cells were then passed through a 20-gauge needle and a 70-μm mesh. Non-specific F_c receptor binding was blocked (TruStain fcX; Biolegend, San Diego, CA) and the cells were then stained for flow cytometry. The conjugated antibodies allophycocyanin (APC)-anti-CD11c (clone: N418), fluorescein isothiocyanate (FITC)- and AlexaFluor (AF)-647-anti-MHC-II (clone: M5/114.15.2) and phycoerythrin (PE)-anti-Siglec-F (clone: E50-2440) (all BD Biosciences, San Jose, CA) were used to identify cDC and macrophages; PE-anti-TCRδ (clone: GL3), PE-Cγ7-anti-CD45 (clone: 30-F11) and FITC-anti-CD3 (clone: 17A2) (all Biolegend) were used to identify γδ T-cells.

To investigate IL-17A production by γδ T-cells, single cell lung suspensions were incubated with or without PMA (phorbol 12-myristate 13-acetate; 100 ng/ml), ionomycin (500 ng/ml) and GolgiStop (BD Biosciences) for 4 h prior to staining. Intracellular IL-17A was detected after permeablization (Cytofix/Cytoperm; BD Biosciences) and staining with APC-anti-IL-17A within γδ T-cells or biotinylated-anti-IL-17A (clone: TC11-18H10.1) with a streptavidin-PE secondary antibody within macrophages. Isotype control antibodies were used to set all gates. Cells were analyzed using a Canto II (BD Biosciences) and the data was analyzed using FlowJo software (Tree Star; Ashland, OR). A minimum of 40 000 events was acquired for each sample.

Quantitative real-time PCR

The right lung was homogenized and total RNA purified using a RNeasy Column kit (Qiagen, Germantown, MD). RNA quantity and purity was determined using a small volume spectrophotometer (Nanodrop; Thermo Scientific, Waltham, MA). A total of 1 μg RNA was converted into cDNA using a Super Script III First-strand amplification kit for qRT-PCR (Life Technologies, Grand Island, NY). Il17a and Il23a mRNA abundances were quantitated by real-time PCR (7300 Real-Time PCR Systems; Applied Biosystems, Carlsbad, CA) using intron-spanning primers (Kasahara et al. 2012; Mathews, Joel, et al. 2014) and SYBR-green detection. Il17a and Il23a were normalized to 36b4 expression using ΔΔCt method (Livak and Schmittgen 2001).

Statistical analysis

Data were analyzed by one-way analysis of variance (ANOVA) followed by a post-hoc Tukey’s t-test, linear regression (Prism 5.0; GraphPad Software, La Jolla, CA) or factorial ANOVA with a post-hoc Fisher-least significant difference test (Statistica; StatSoft, Tulsa, OK). Non-normally distributed data was log transformed prior to statistical analysis. A p-value < 0.05 was considered statistically significant.

Results

IL-23 deficiency reduces O₃-induced IL-17A expression

Compared to air, O₃ (0.3 ppm for 72 h) caused a significant increase in both pulmonary Il23a and Il17a mRNA abundances (Figure 1), as noted earlier. To determine whether IL-23 contributed to induction of Il17a expression, pulmonary Il17a mRNA abundance (BAL IL-17A was below limit of detection of ELISA) was measured in WT and IL-23^−/− mice exposed to O₃. Il17a mRNA abundance was significantly reduced in IL-23^−/− vs in WT mice (Figure 2A).

Figure 1. (A) Pulmonary Il23a and (B) Il17a mRNA abundance, determined by qRT-PCR, in WT mice exposed to room air or ozone (0.3 ppm) for 72 h. Results shown are means ± SE from 5–6 mice/group and are expressed relative to the air exposed mice. *p < 0.05 vs air.

Figure 2. (A) Pulmonary Il17a mRNA abundance (expressed relative to in WT mice). (B) Pulmonary IL-17A⁺ γδ T-cells. (C) Pulmonary IL-17A⁺ interstitial macrophages measured by flow cytometry in WT and IL-23^−/− mice exposed to ozone (0.3 ppm for 72 h). Results shown are means ± SEM of 8–9 mice/group for qRT-PCR and four mice/group for flow cytometry. *p < 0.05 vs WT mice.

Earlier studies in this laboratory noted that both interstitial macrophages and γδ T-cells express IL-17A after O₃ exposure. To determine which cell type was impacted by an IL-23 deficiency, flow cytometry was used to quantify IL-17A expressing macrophages and γδ T-cells in the lungs of WT and IL-23^−/− mice exposed to O₃. The total number of γδ T-cells was not different in O₃-exposed WT and IL-23^−/− (data not shown), nor was there any difference in the number of IL-17A⁺ γδ T-cells (Figure 2B). In contrast, IL-17A⁺ macrophages (Figure 2C) were lower in the lungs of the IL-23^−/− vs in those of the WT mice.

Despite the significant reduction in Il17a mRNA expression in IL-23^−/− vs in WT mice (Figure 2A) and data indicating that IL-17A is required for PMN recruitment after subacute O₃ exposure (Kasahara et al. 2012; Mathews et al. 2013), we observed only a non-significant (≈25%) reduction in BAL PMN in O₃-exposed IL-23^−/− vs WT mice (Figure 3A), nor was there any change in BAL macrophage numbers (Figure 3B). Similarly, IL-23 deficiency resulted in a reduction in BAL G-CSF concentrations (86 [± 40] vs 52 [± 12] pg/ml in WT and IL-23^−/− mice, respectively), but the effect was not significant. There were also no significant IL-23-dependent changes in BAL concentrations of other cytokines that either are involved (IL-6, CXCL5, CXCL1, CXCL2, IL-1) or might be involved (CCL3, CCL5) in neutrophil recruitment to the lungs after O₃ exposure (data not shown). Nevertheless, IL17a mRNA expression did appear to account for at least part of the variance in BAL PMN among WT mice, since there was a significant inverse correlation (r² = 0.444, p < 0.05) between the percentage of BAL cells that were PMN and Il17a mRNA abundance (Figure 3C) in WT mice (note that an increase in ΔΔCt indicated reduced Il17a expression). In contrast, no such correlation existed in the IL-23^−/− mice (r² = 0.005, NS) (Figure 3C).

Figure 3. Bronchoalveolar lavage (BAL) (A) neutrophils and (B) macrophages in WT and IL-23^−/− mice exposed to ozone (0.3 ppm for 72 h). Results shown are means ± SE of 8–9 mice/group. (C) Log %BAL neutrophils plotted vs ΔΔCt values for Il17a. All mice were ozone-exposed. The regression line shown was calculated from the combined data from the WT and IL-23^−/− mice. Note: increase in ΔΔCt indicates reduced Il17a expression.

IL-1 contributes to O₃-induced PMN recruitment via an IL-17A-/IL-23-independent mechanism

Compared to saline, there was no effect of the IL-1R1 antagonist anakinra on either Il17a or Il23 expression in O₃-exposed mice (Figure 4A and B). Nevertheless, O₃-induced increases in BAL PMN numbers were significantly lower in anakinra- vs saline-treated mice (Figure 4C), although macrophage numbers and BAL protein concentrations (a measure of O₃-induced lung injury) were not affected (Figures 4D and 4E). A multiplex assay was used to measure other IL-1-dependent cytokines and chemokines that could be contributing to the changes in PMN numbers. The data shown in Figure 5 illustrate that IL-6, LIF and GM-CSF concentrations were each significantly reduced in anakinra-treated vs in saline-treated mice exposed to O₃.

Figure 4. Pulmonary (A) Il17a and (B) Il23a mRNA abundance (expressed relative to WT mice) and BAL (C) neutrophils, (D) macrophages and (E) total protein in WT mice exposed to ozone (0.3 ppm for 72 h) that had been treated with anakinra or saline. Results shown are means ± SE of 6–12 mice/group. *p < 0.05 vs saline-treated mice.

Figure 5. BAL (A) LIF, (B) IL-6 and (C) GM-CSF assayed by multiplex from WT (C57BL/6J) mice exposed to ozone (0.3 ppm for 72 h) that had been treated with anakinra or saline. Results shown are means ± SE of eight mice/group. * p < 0.05 vs saline-treated mice.

Conventional dendritic cells do not contribute to O₃-induced PMN recruitment to the lungs

Although other O₃ exposure paradigms have been shown to increase pulmonary DC numbers (Brand et al. 2012), the effects of subacute O₃ (0.3 ppm for 24–72 h) have not been established. Consequently, flow cytometry was used here to quantify the number of DC in lungs of mice exposed to room air or to O₃ for 24, 48 or 72 h. DC (CD11c⁺MHCII⁺Siglec-F⁻ cells) (Desch et al. 2014) were identified using the gating strategy shown in Figure 6(A). In WT mice, O₃ caused a time-dependent increase in the number of pulmonary DC (Figure 6B): DC counts were not different in air-exposed mice and mice exposed to O₃ for 24 h, but O₃ exposure for 48 or 72 h did increase the number of pulmonary DC.

Figure 6. (A) Gating strategy used to assess lung DC. Cells were first gated based on forward scatter (FSC) and side-scatter (SSC) characteristics. Cells positive for MCH-II and CD11c were determined to be either dendritic cells (DC, SiglecF⁻) or macrophages M∅, SiglecF⁺). (B) Total number of MHC-II⁺/CD11c⁺/SiglecF⁻ DC in WT mice exposed to air or ozone (0.3 ppm) for 24, 48 or 72 h. (C) Total number of MHC-II⁺/CD11c⁺/SiglecF⁻ DC in WT and Flt3l^−/− mice exposed to air or ozone (0.3 ppm) for 72 h. Results shown are means ± SE of 7–8 mice/group. *p < 0.05 vs air-exposed mice with same genotype; ^#p < 0.05 vs WT mice with same exposure.

To determine whether cDC contributed to pulmonary Il23a and Il17a expression after O₃ exposure, Flt3l^−/− mice which lack cDCs due to a requirement for Flt3l in differentiation of the common myeloid progenitor (Waskow et al. 2008) were utilized. Indeed, DC were significantly lower in the lungs of Flt3l^−/− than in WT mice and there was no O₃-induced increase in DC numbers in the Flt3l^−/− mice (Figure 6C). There was no difference in pulmonary Il23a or Il17a mRNA abundance in O₃-exposed Flt3l^−/− vs in WT mice (Figures 7A and B), although there was significantly more Il17a in the lungs of Flt3l^−/− vs in WT mice exposed to air. The O₃-induced increases in BAL PMN numbers were also unaffected by Flt3l genotype (Figure 7C).

Figure 7. Pulmonary (A) Il23a and (B) Il17a mRNA abundance (expressed relative to WT air mice) and (C) BAL neutrophil numbers in WT and Flt3l^−/− mice exposed to room air or ozone (0.3 ppm) for 72 h. Results shown are means ± SE of 7–8 mice/group. *p < 0.05 vs air-exposed mice with same genotype; ^#p < 0.05 vs WT mice with same exposure.

Discussion

The data here indicate that IL-23 contributes to the IL-17A expression induced by subacute O₃ exposure, at least in part via effects on lung interstitial macrophages. The data also indicate that IL-1 signaling contributes to the PMN recruitment induced by subacute O₃ in an IL-17A- and IL-23-independent manner. Finally, the data demonstrate that, despite a robust increase in cDC numbers after O₃ exposure, and the ability of these cells to express IL-23 in response to other stimuli (Fitch et al. 2007), cDCs are not the source of IL-23-dependent Il17a expression.

In O₃-exposed mice, IL-23 deficiency caused a significant reduction in pulmonary Il17a mRNA abundance. These data were consistent with observations that Il23a expression is induced by subacute O₃ exposure (Mathews, Williams, et al. 2014) and with data of other investigators showing that IL-23 promotes pulmonary IL-17A expression induced by other stimuli, including infection with Streptococcus pneumonia or Pseudomonas aeruginosa (Dubin et al. 2012; Kim et al. 2013). However, while Il17a mRNA was substantially reduced (≈70% reduction) in IL-23^−/− vs WT mice (Figure 2A), it was not completely abolished. Similarly, Ghilardi et al. (2004) observed only a partial reduction in T-cell IL-17 expression after in vitro stimulation. The results indicated that, while IL-23 promoted pulmonary IL-17A production after subacute O₃ exposure, other factors that drive IL-17A were also important. For example, serum amyloid A, TGFβ, IL-6, IL-1 (Kim et al. 2013) and tumor necrosis factor (TNF)-α (Mathews, Williams, et al. 2014) can each induce IL-17A in various cell types. We previously reported no effect of IL-6 deficiency on IL-17A expression in WT mice after subacute O₃ exposure, although IL-6 does contribute in adiponectin-deficient mice in which O₃ induced much higher concentrations of IL-6 (Kasahara et al. 2014). (See additional discussion of other IL-17A promoting cytokines below.)

Flow cytometry was performed to determine which cells capable of expressing IL-17A were responsible for the IL-23 dependent changes in Il17a mRNA (Figure 2). Subacute O₃ exposure increases the numbers of IL-17A⁺ CD11c⁻ macrophages and IL-17A⁺ γδ T-cells in the lungs (Kasahara et al. 2012; Mathews, Williams, et al. 2014). Both cell types express the IL-23R and the IL-12RB1 proteins necessary for responses to IL-23 (Parham et al. 2002). IL-23 can induce IL-17A expression in γδ T-cells (Lockhart et al. 2006; Sutton et al. 2009) in some settings. However, in a mouse model of colitis, IL-23 is not required for production of IL-17A from colonic γδ T-cells (Lee et al. 2015). Similarly, no reductions in IL-17⁺ γδ T-cells were observed in IL-23^−/− mice exposed to O₃, despite substantial reductions in Il17a mRNA expression in these mice. Instead, TNFα appears to contribute to expression of IL-17A in γδ T-cells after subacute O₃.

IL-17A⁺ γδ T-cells were reduced ≈ 80% in TNFR2^−/− vs in WT mice and the TNFα antagonist, etanercept, also caused substantial reductions in Il17a mRNA abundance in WT mice after subacute O₃ exposure (Mathews, Williams, et al. 2014). In contrast, IL-17A⁺ macrophages were reduced in IL-23^−/− mice. Similarly, IL-23 has been shown to induce IL-17A release from peripheral blood mononuclear cells from patients with atherosclerosis (Abbas et al. 2015). Macrophages also have the capacity to produce IL-23 (Bosmann et al. 2013; Abbas et al. 2015), suggesting these cells may use autocrine IL-23 signaling to promote IL-17A production. However, even IL-17A⁺ macrophages were reduced by only 50% in IL-23^−/− mice and IL-17⁺ γδ T cells were unaffected, despite ∼70% reductions in Il17a mRNA abundance in these mice (Figure 2). The results suggest that other unexamined cells may also contribute to IL-17A expression after O₃ exposure. We have previously reported that T_H17 cells are not involved (Kasahara et al. 2012), but innate lymphoid cells type 3 (ILC3) that also produce IL-17A are found in the lung in another condition of oxidative stress, i.e. obesity (Kim et al. 2014) and could be contributing to O₃-induced increases in Il17a expression.

We have previously reported that, compared to isotype control antibody, anti-IL-17A causes a substantial reduction in BAL PMN numbers after subacute O₃ exposure in mice (Mathews, Williams, et al. 2014). Hence, it was somewhat surprising that, despite ≈70% reduction in Il17a mRNA abundance in IL-23^−/− vs in WT mice exposed to O₃, there was no significant reduction in BAL PMN counts. The lack of effect of IL-23 deficiency on G-CSF and other cytokines and chemokines known to be required for O₃-induced PMN recruitment (e.g. IL-6 and CXCL1) is consistent with the lack of effect of IL-23 deficiency on BAL PMN. Given the very marked increase in l17a mRNA abundance with O₃ (≈60-fold) seen here, even with a 70% reduction in Il17a expression induced by IL-23 deficiency there would still be substantial Il17a expression in the IL-23^−/− mice, which may be why PMN were only minimally affected. Additionally, increases in pulmonary Il23a mRNA abundance were not observed until 48 h of O₃ exposure (Mathews, Williams, et al. 2014). In contrast, increases in pulmonary Il17a mRNA abundance were already observed after only 24 h of exposure, as were increases in IL-17A⁺ cell numbers, although Il17a continued to increase up to 72 h of exposure. Recruitment of BAL PMN to the lung also begins as early as 24 h after initiation of exposure (Mathews, Williams, et al. 2014). Taken together, the data here suggest to us that, after initiation of O₃ exposure, factors other than IL-23 contribute to the early induction of IL-17A (likely from γδ T-cells) and consequent PMN recruitment, and that IL-23 induces additional IL-17A expression, likely from macrophages, only later in the exposure.

IL-1α and IL-1β are both expressed in the lungs after subacute O₃ exposure (Johnston et al. 2007). Furthermore, IL-1R1 signaling has been shown to induce IL-23 expression in myeloid cells in response to other stimuli (Andersson et al. 2004). IL-1R1 signaling can also induce IL-17A expression in a variety of cell types (Chung et al. 2009; Gasse et al. 2011; Kim et al. 2014). Nevertheless, in the present study, there was no effect of the IL-1R1 antagonist anakinra on O₃-induced increases in either pulmonary Il23a or Il17a mRNA abundance. However, given that anakinra was only administered 24 h prior to the initiation of O₃ exposure, we cannot rule out the possibility that more sustained IL-1 inhibition might have been effective. In this context, it is conceivable the reduced Il17a mRNA abundance observed in the O₃-exposed IL-23^−/− mice might have been the result of developmental effects of IL-23 that promoted the ability of macrophages and other cells to subsequently produce IL-17A.

Whereas anakinra did not affect either Il17a or Il23a abundance in O₃ exposed mice, it caused a substantial reduction in BAL PMN counts. The reduction in BAL PMN counts was consistent with previous observations of reduced BAL PMN numbers in IL-1R1^−/− vs in WT mice exposed to subacute O₃ (Johnston et al. 2007) and was likely a consequence of anakinra-induced reductions in BAL concentrations of IL-6, LIF and GM-CSF (Figure 5). Both IL-6 and LIF can induce PMN recruitment and activation through binding of gp130 to their respective receptors (Murakami et al. 1993). Indeed, we previously reported reduced BAL PMN levels in IL-6^−/− vs in WT mice after subacute O₃ exposure (Johnston et al. 2005; Lang et al. 2008; Kasahara et al. 2014). GM-CSF also has the capacity to regulate chemotaxis and survival of PMN (Laan et al. 2003; Khajah et al. 2011) and IL-1β can induce the production of GM-CSF within the lung (Lukens et al. 2012).cDC have the capacity to release IL-23 during inflammation induced by a variety of stimuli, including cytokines, DAMPs and microbes (reviewed by Lambrecht and Hammad 2009). Nevertheless, no significant reduction in Il23a or Il17a mRNA abundance in Flt3l^−/− vs in WT mice was observed here despite major reductions in the number of pulmonary DC in the Flt3l^−/− mice. BAL PMN counts were also no different in Flt3l^−/− vs in WT mice. Flt3l^−/− mice are known to have a reduced white blood cell count, but this difference is largely attributed to a reduction in circulating lymphocytes rather than PMN (McKenna et al. 2000). The results indicated that subacute O₃ exposure did not provide the stimuli necessary for induction of IL-23 in cDC. Indeed, studies in which IL-23 has been observed in cDC typically used antigenic stimulation (e.g. bacteria, autoantibodies) (Roses et al. 2008; Sutton et al. 2009). Instead, the IL-23 produced after subacute O₃ exposure here likely originated from other myeloid cells rather than from cDC.

Conclusion

The data from the present study indicate that IL-23 is required, at least in part, for increases in pulmonary IL-17A expression that occur after subacute O₃ exposure in mice and that one target of IL-23 is likely IL-17A⁺ interstitial macrophages. The data also indicate that O₃-induced increases in pulmonary expression of IL-17A and IL-23 do not require IL-1, but that IL-1 nevertheless contributes to O₃-induced neutrophil recruitment to the lungs, likely via effects on IL-6, LIF, and GM-CSF. Finally, this study showed that, despite increases in pulmonary cDC after subacute O₃, these cells do not contribute to the IL-17A and IL-23 expression or the neutrophil recruitment induced by O₃.

Acknowledgments

The authors would like to thank Nico Ghilardi (Genentech, San Francisco, CA) for providing IL-23^−/− breeding pairs. This study was supported by ES013307 and ES000002. JAM was supported by ES02256. JDB was supported by HL007118.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.