A prominent air pollutant, Indeno[1,2,3-cd]pyrene, enhances allergic lung inflammation via aryl hydrocarbon receptor

Chronic exposure to ambient polycyclic aromatic hydrocarbons (PAHs) is associated with asthma, but its regulatory mechanisms remain incompletely defined. We report herein that elevated levels of urinary 1-hydroxypyrene, a biomarker of PAH exposure, were found in asthmatic subjects (n = 39) as compared to those in healthy subjects (n = 43) living in an industrial city of Taiwan, where indeno[1,2,3-cd]pyrene (IP) was found to be a prominent PAH associated with ambient PM2.5. In a mouse model, intranasal exposure of mice with varying doses of IP significantly enhanced antigen-induced allergic inflammation, including increased airway eosinophilia, Th2 cytokines, including IL-4 and IL-5, as well as antigen-specific IgE level, which was absent in dendritic cell (DC)-specific aryl hydrocarbon receptor (AhR)-null mice. Mechanistically, IP treatment significantly altered DC’s function, including increased level of pro-inflammatory IL-6 and decreased generation of anti-inflammatory IL-10. The IP’s effect was lost in DCs from mice carrying an AhR-mutant allele. Taken together, these results suggest that chronic exposure to environmental PAHs may pose a significant risk for asthma, in which IP, a prominent ambient PAH in Taiwan, was shown to enhance the severity of allergic lung inflammation in mice through, at least in part, its ability in modulating DC’s function in an AhR-dependent manner.


Results
IP was a prominent PAH associated with ambient PM 2.5 . After monthly samplings of atmospheric PM 2.5 in 4 monitoring stations in a subtropical industrial city Kaohsiung, 16 PAHs (USEPA priority) were analyzed. Results showed that among the 16 PAHs analyzed, IP was a prominent PAH associated with ambient PM 2.5 in all sites (Fig. 1A) and all seasons (data not shown). The concentrations of PAHs were higher in winter and autumn than in summer (Fig. 1B). It is also noted that the average concentrations of IP was around four times higher than that of BaP in all sites and all seasons (Fig. 1C). In a case-control design, the urinary concentration of OH-Py, a major metabolite of pyrene and widely used as a marker of exposure to PAHs, was significantly higher in subjects with asthma than those in control subjects (mean ± SEM, 0.092 ± 0.011 μg/g creatinine (CRN) versus 0.043 ± 0.003 μg/g CRN, P < 0.0001).

IP exposure potentiated antigen-induced pulmonary allergic inflammation. To determine
whether IP was able to modulate airway allergic inflammation, a murine model was designed to mimic human exposure. The results showed that intranasal exposure of IP at concentrations ranging between 0.4 -2 μM significantly enhanced the severity of pulmonary allergic inflammation, including increased infiltration of eosinophils, lymphocytes, and neutrophils ( Fig. 2A), as well as enhanced IL-4 and IL-5 levels in bronchoalveolar lavage fluids (BALFs) (Fig. 2B). In addition, the levels of serum ovalbumin (OVA)-specific IgE were significantly enhanced in mice exposed to IP in a dose-dependent manner (Fig. 2C). These data suggest that IP may act as an adjuvant to modulate the allergic inflammation.
In order to ensure the adjuvant role of IP in allergic responses, we tested the response induced by IP alone without antigen immunization. As shown in Fig. 3, IP exposure alone did not induce any allergic response, which was similar to vehicle control group (no OVA and no IP treatment). However, in the presence of OVA, IP exposure significantly enhanced OVA-induced allergic responses, including the levels of immune cell infiltration (Fig. 3A), Th2 cytokines (Fig. 3B) and OVA-specific IgE Abs (Fig. 3C). Further, immunocytochemistry analysis showed that while IP exposure alone did not induce apparent airway inflammation, prominent inflammatory lesions were noted in IP-treated, OVA-sensitized and challenged mice (Fig. 3D). These data suggest that airway IP exposure may act as an adjuvant to modulate immune responses towards Th2 responses, leading to the enhanced severity of allergic asthma. IP exposure modulated DC's function in an AhR-dependent manner. IP has similar structure as BaP, which is a well-known ligand for AhR. Thus, we examined whether AhR was involved in IP-mediated effect on allergic lung inflammation. As shown in Fig. 4, administration of a commonly used AhR antagonist, CH223191, significantly inhibited the IP's effect on OVA-induced influx of eosinophils and lymphocytes, Th2 cytokines, and OVA-specific IgE levels. Because DC plays an essential role in the pathogenesis of allergic asthma, we then asked whether IP modulated DC's function through AhR by the use of C57BL/6 mice carrying a high-responder allele (AhR b-1 ) and a congenic strain carrying a low-responder allele of AhR (AhR d ). The affinity of AhR d for its ligand is 10-100 times lower than of AhR b-1 allele 25 . As shown in Fig. 5A and B, IP treatment at relatively low doses starting from 0.1 to 100 nM significantly inhibited LPS-induced expression of CD40, CD80 and IL-10 production, but enhanced LPS-stimulated IL-6 secretion in BM-DCs. However, IP exposure lost its effect on BM-DCs from AhR d mice; also, IP exposure did not affect the allergic parameters in OVA-immunized DC-specific AhR-null (DC-AhR −/− ) mice compared to those in wild type mice (Fig. 5C-F). Taken together, these data suggest that IP exposure may disturb DC's function in an AhR-dependent manner, leading to the enhanced allergic lung inflammation.
It has been demonstrated that ultrafine particles associated with PAHs can induce oxidative stress and cytotoxicity in macrophages and epithelial cells 6 . Thus, we analyzed whether IP treatment promoted the production of 4-HNE, a sensitive oxidative stress marker, in the asthmatic murine model. As shown in Fig. 6, IP treatment in conjunction with OVA did enhance the 4-HNE generation in lung tissues as compared to those seen with IP or OVA treatment alone. Interesting, IP treatment alone without antigen challenge in mice also showed more 4-HNE-positive staining in lung tissues as compared to those in vehicle control group (no IP, no OVA treatment). Further, the level of 4-HNE in serum from OVA immunized mice was significantly increased upon IP exposure in a dose-dependent manner.

Discussion
PAHs associated with air pollution have been considered as an important environmental risk for asthma 26 ; however, the responsible component(s) of PAHs and the involved mechanism have not been well defined. In the present study, we identified IP as the main component of ambient PAHs in southern Taiwan, and showed its potential ability to enhance the severity of allergic lung inflammation through modulating DC's function in an AhR-dependent manner. Further, biomonitoring analysis of a case-control study population showed that the urinary levels of OH-Py, a surrogate marker for PAH exposure, were significantly higher in patients with asthma than those seen in controls. This is the first study, to our knowledge, in providing evidence supporting a causal relationship between airborne IP exposure and lung inflammation using a mouse model of asthma.
The sources of IP were identified mainly from traffic 27 . However, as Kaohsiung City is an industrial city with steel plants, coal and coke combustion should also be considered as a source of the presence of IP in PM 2.5 7 . Other atmospheric exposure of IP could include cooking related activities. For example, the concentration of IP in kitchen indoor air was found to be above the permissible Occupational Safety and Health Administration (OSHA) norms for indoor air 28 . Even though IP was found present in relatively higher amounts in many environmental matrices, the study of its toxicity has been scarce. While the toxic equivalent factor (TEF) with respect to BaP of IP has been assessed as 0.1 29 , the functional effect of IP on DCs appears to be more potent than that of BaP. BaP has been shown to inhibit DC maturation and IL-12 secretion in murine BM-DCs 30 and in human monocyte-derived DCs 20 at the concentrations of 1 and 10 μM, respectively. However, in our study, IP significantly suppressed IL-10 production and maturation in murine BM-DCs at the concentration of 0.1 nM and 100 nM, respectively (Fig. 5A,B). In addition, at least during the period between 2012 and 2013 in southern Taiwan, the amount of IP was around 4-fold higher than BaP in ambient air PAHs (Fig. 1C). It suggests, therefore, that IP, in addition to BaP, among PAHs associated with PM 2.5 , may contribute to the impact of air pollution on human health.
It has been noted that two prototypical AhR ligands, TCDD and FICZ, were shown to be able to modulate DCs with tolerogenic phenotypes and facilitate Foxp3 + regulatory T cell differentiation at relatively higher doses 22 . However, in our study, chronic exposure of IP at human exposure dosages in mice showed enhanced levels of pulmonary allergic inflammation. While it is currently unclear as to the reason why AhR's ligands mediate these seemingly dichotomized immune responses, it may be due, in part, to a potential ligand-specific effect or, more than likely, different doses used in various studies. Consistent with our findings, Zhou et al. used relatively low-doses of AhR ligands, TCDD and FICZ, and showed enhanced mast cell response, including degranulation and cytokine generation 31 , while at higher doses following repetitive administration, FICZ (100 μg/kg/mouse) may have an inhibitory effect on mast cell's response 32 , It is likely, therefore, that low doses of AhR's ligand exposure may mediate immune activation, rather than tolerance. In fact, it is known that AhR ligand's effect does not conform to the monotonic response curve. Nevertheless, further studies are needed to define the molecular basis of the AhR-ligand axis and its impact on immune system. 4-HNE is a reactive lipid mediator, which is generated from lipid peroxidation. Due to its electrophilic property, 4-HNE can form protein as well as lipid adducts and then modulate a number of signaling processes and impact cellular functions (reviewed in reference 33 ). In the present study, 4-HNE was observed in the lung tissue of intranasally IP treated alone mice, suggesting exposure to PAHs may result in oxidative stress in the lungs. Although IP alone treated mice did not show obvious local allergic inflammation, pre-exposure to IP through airway significantly enhanced the severity of OVA-induced lung inflammation. These data suggest that exposure to environmental PAHs at human relevant exposure ways and doses is not expected to generate a dramatic "toxic" effect, but rather it may persistently modify the level of the host response. As a corollary, a recent study by Perzanowski et al. showed that prenatal exposure to cockroach allergen was associated with a greater risk of allergic sensitization in children and this risk was enhanced by exposure to PAHs 34 . These data support the link between air pollution and the severity of asthma in humans 35 .
Our study demonstrated that IP may act through AhR to alter DC function, and lead to enhanced allergic inflammation. As shown in Fig. 4, AhR antagonist CH223191 could reverse the IP effect on allergic lung inflammation, including inflammatory cell infiltration in BALFs and tissue, Th2 cytokine production and OVA-specific IgE levels. In addition to conventional DCs, several AhR-expressing cell types may also contribute to the development of IP-enhanced allergic inflammation in airway, such as epithelial cells 36 , fibroblast 37 , innate lymphoid cells 38 , mast cells 15,31 as well as pDCs 39 . However, conventional DCs seem to play the major role in the IP's effect because chronic airway exposure to IP in DC-AhR −/− mice did not enhance the allergic lung inflammation in the OVA-induced asthma model. As IL-10 is a well-characterized immune suppressive cytokine, decreased expression of IL-10 in DCs may contribute to the IP's effect on allergic inflammation. Moreover, two likely, but nor exclusive, mechanisms could be operated in IP's potentiating effect. One is that airway DCs can be directly and chronically exposed to IP, which results in alteration of DC's function as they carry the antigen and are capable of modulating the immune response via trafficking between epithelial cells and draining lymph nodes 40,41 . The other possible mechanism is that IP firstly acts on airway epithelium, which induces epithelial cells to release mediators and modulate local DC function. This evidence was provided by our cooperative study by Wang et al. 42 . The authors demonstrated that intranasal exposure to IP was sufficient to induce the generation of a lipid mediator, sphingosine-1-phosphate, from either mast cells, or epithelial cells or the combination of both without allergen sensitization in a mouse model, while sphingosine-1-phosphate is a potent bioactive lipid mediator that regulates lymphocyte trafficking, cell growth, apoptosis, inflammation and, perhaps, DCs 43 . In support of this notion, we also found that intranasal IP exposure alone without any allergen sensitization resulted in oxidative stress response in the lungs (Fig. 6A), but not inflammation (Fig. 3D), and as the consequence, significantly enhanced severity of OVA-induced airway inflammation was noted (Fig. 3). This suggests that PAHs may potentiate the development of asthma through the AhR-ligand axis in priming immune response. However, the detailed mechanism regarding how IP-AhR axis down-regulates IL-10 expression in DCs and its detailed mechanisms on the regulation of allergic asthma await further in-depth studies.
In the present study, asthmatic subjects in southern Taiwan showed elevated environmental PAH exposure. Chronic exposure to IP, a prominent ambient PAH, may affect DC functions through modulating AhR activity, leading to enhanced allergic inflammation in an established asthma murine model. Our study provides evidence supporting a direct link between ambient PAH exposure and the expression of allergic diseases.

Methods
Air sampling and PAH detection. One year air sampling campaign was carried out in four sampling sites in Kaohsiung City, Taiwan. The details of sampling and measurement of PAHs in PM 2.5 are presented in reference 27 . Briefly, the 24-hr PM 2.5 samples (twice a month, on 200 mm × 250 mm quartz fiber filter) were collected sequentially at each of the four sampling sites during the period of December 2012 to November 2013. The four sampling sites are SG, located in southern part of Kaohsiung with petrochemical, shipbuilding, firepower plant, iron, and steel industrial activities; FS, located in the northern part with petrochemical industrial activities nearby and densely populated; AG, located in the center of the city with intense traffic and commercial activities; while NSYSU is a coastal area site.
For PAH analysis, the quartz fiber filters were extracted with dichloromethane using an accelerated solvent extractor (Dionex ASE 300). The extracts were concentrated to 0.5 mL for GC-MS, a capillary gas chromatograph (Agilent 6890 N) and a mass spectrometer (Agilent 5973 N) Patients. The study population included adult patients with current asthma (n = 39) at the outpatient departments of 8 medical centers following established protocols. We also enrolled normal control subjects (n = 43)  Bone marrow cells from C57BL/6 and AhR d mice were treated with IP starting from day 1 to day 8, and LPS was added for another 24 hrs. The BM-DCs were harvested for phenotypic analysis (A) and the levels of cytokines in the culture supernatants were measured by ELISA (B). Data shows mean ± SD of three independent experiments. C57BL/6 (WT) or DC specific AhR-null mice (DC-AhR −/− ) were intranasally pre-exposed to vehicle (NC) or IP (2 μM) and then immunized with OVA. All mice were challenged with OVA aerosol for three consecutive days. Cell subsets (C) and cytokine levels (D) in BALFs were both determined by flow cytometry. The line within the vertical points marks the mean for each group. (E) Serum levels (mean ± SEM) of OVA-specific IgE Abs were measured by ELISA. n = 7-10/group. (F) Representative lung sections stained with hematoxylin and eosin. The magnification: 200×. *P < 0.05 vs. control (Mann-Whitney U test). Scale bars represent 50 µm in (F). selected from volunteers and from those who requested annual physical examinations without respiratory diseases. All subjects were non-smokers. All eligible subjects were enrolled in the study after signing the informed consent approved by the respective recruitment hospitals. Patients with age ≥18 years were enrolled and met the following inclusion criteria: (1) at least 18 years of age, (2) physician-diagnosed asthma. Physician's diagnosis of asthma was made according to the 2008 Global Initiative for Asthma (GINA) guidelines. In this study, we further stratified asthma into four groups: asthma that required GINA Step 1 or 2 to maintain good control was considered as mild asthma, GINA Step 3 as moderate, GINA Step 4 as severe (Table 1). Patients with severe asthma (step 4) were receiving more than two combination controller therapies (ICS, LABA, leukotriene modifier and sustained-release theophylline). Patients were also evaluated for their control status by using asthma control test (ACT), a validated patient-completed questionnaire consisting of five parameters aimed at assessing asthma symptoms (daytime and nocturnal), use of rescue medications, and the effect of asthma on daily functioning. The scores range from 5 (poor control of asthma) to 25 (complete control of asthma). The scores equal or less than 19 was considered to be "not well controlled". Pulmonary function was measured with a Jaeger Master screen Pulmonary System spirometer (Hoechberg, Germany). At baseline, FEV1 and FVC were expressed as percentages of the predicted values, whereas FEV1/FVC was reported only as a ratio. The smoking status of all patients was also assessed and divided into three groups: current smokers, ex-smokers who had stopped smoking for a minimum of 1 month before the initial visit, and lifetime non-smokers. Patients have one of self-reported allergic triggers (pollen, mold, dust, animals, beans, seafood, milk and eggs), allergic rhinitis and/or atopic dermatitis were defined as having positive history of allergy. Urine samples were collected and stored frozen in glass container until analysis. In this dataset, demographical comparison revealed significant differences in second-hand smoke at home and lung function parameters (FEV1 and FEV1/FVC ratio) between controls and patients ( Table 1). The study protocol was approved by the Institutional Review Boards of Kaohsiung Medical University Hospital (KMU-IRB-990392). The study was performed in accordance with the ethical standards laid down in the Declaration of Helsinki.
Urinary PAH metabolite detection. Urinary OH-Py was measured with previously published assay 44 and used as a surrogate marker for the exposure of PAHs, as the standard IP metabolite was not commercially available at the time of analysis. First, the defrosted urine (10 mL) in acetate buffer (0.1 M acetic acid and 0.1 M sodium acetate) was digested with α-glucuronidase/aryl-sulphatase, and then OH-Py in the urine samples was extracted, enriched and purified by cartridges packed with C-18 reversed-phase liquid chromatographic material (Waters SEP-PAK VAC C-18, Waters, Milford, MA, USA) at a loading rate of less than 3 mL/min. The retained solutes in cartridges were eluted with 6 mL of isopropyl alcohol and dried. The processed sample was dissolved in 2 mL of isopropyl alcohol for analysis, using a high-performance liquid chromatography (Waters 2695) with fluorescence detector (Waters 474) and a 150 mm × 4 mm LiChrosorb RP-18 (5 mm) column (Supelo). The mobile phase was 65% methanol (methanol/water 1⁄4 65:35). The excitation wavelength and emission wavelength of fluorescence were 281 and 388 nm, respectively. This method had a limit of detection of 0.028 ng/mL, with an average recovery rate of 96.18%, and a coefficient of variance smaller than 10% for repeated measurements. The urinary levels of OH-Py were normalized by the levels of urinary CRN and expressed as μg/g CRN. Statistical Analysis. The independent t-tests were used to examine the differences of age, body mass index, and urine CRN. Chi-square tests were used to compare the difference of gender, smoking habits, and second-hand smoke status. One-way ANOVA was used to compare the difference among groups. Nonparametric Mann-Whitney U tests were conducted to examine the differences of human urinary OH-Py and the results from control and treated cells or mice. P values < 0.05 were considered significant. All statistical tests were performed by SPSS for Windows, version 22.0. (SPSS Inc., Chicago, Ill., USA).