This study subjects were from Taiwanese Maternal and Infant Cohort pilot study, which was approved by the Institutional Review Board of E-Da Hospital (EMRP35101N) and Kaohsiung Medical University Hospital (KMUHIRB-2012-11-02(I)). Written informed consent was obtained from all pregnant women in the study and on behalf of their children. The detailed study design and questionnaire are comprehensively described elsewhere (12). Children were evaluated at the age of 3 years to investigate allergic symptoms. Physician diagnosis of asthma, atopic dermatitis, or allergic rhinitis were made according to an operational description suggested by the Global Initiative for Asthma (GINA) guidelines and the Allergic Rhinitis and Its Impact on Asthma (ARIA) guidelines. The exclusion criteria for children included severe systemic diseases, such as congenital heart diseases, or autoimmune diseases.

Urinary DEHP metabolites and creatinine levels were measured in single urine sample from the pregnant women according to a modified method previously described (12). Creatinine-corrected adjustment for each individual was used to examine associations between the compound exposure and allergic diseases. The daily intake level of DEHP (∑DEHP) was estimated using a creatinine excretion-based model, and the detailed method was described in this cohort profile (13).

Human bronchial epithelium (HBE) cells (American Type Culture Collection, Rockville, USA) were cultured in keratinocyte-SFM (Thermo Fisher Scientific, Waltham, USA) supplemented with 25 nM hydrocortisone and an 850 nM insulin solution (Sigma–Aldrich, St. Louis, USA). The A549 human lung carcinoma cell line (American Type Culture Collection) was cultured in MEM supplemented with 10% fetal bovine serum, 1% nonessential amino acids, 1% sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin. Cells were centrifuged, resuspended in fresh media, plated in 6-well plates at a density of 1 × 10⁶ cells/ml, and incubated for 24 h before use in experiments. The cells were treated with DEHP or mono-(2-ethylhexyl) phthalate (MEHP) (Sigma–Aldrich) at different concentrations (0.001-0.1 μM) or combined with house dust mites (HDM) extract (Greer Laboratories, Lenoir, USA), which experimental dose of HDM was referred to previous study (14) for 3 or 24 h for isolated RNA or protein measurement. Cells were pretreated with CH-223191 (AhR antagonist, Sigma–Aldrich), flavonoids (apigenin or kaempferol, Sigma–Aldrich), the histone acetyltransferase inhibitor, anacardic acid (AA, Sigma–Aldrich), or the methyltransferase inhibitor, 5’-deoxy-5’-(methylthio) adenosine (MTA, Sigma–Aldrich) 1 h before DEHP or MEHP treatment to identify involved receptors or signaling pathways. The experimental dose of phthalates, AhR antagonist, flavonoids and histone acetyltransferase/methyltransferase inhibitor were referred to our previous publication (15).

Total RNA and cDNA was performed as described in our previously published study (16). qRT–PCR was performed using SYBR Green PCR Master Mix on an ABI 7500 Real-Time PCR system. The qRT–PCR primers for IL33 are as follows: forward: 5’-caaagaagtttgccccatgt, reverse: 5’-aaggcaaagcactccacagt according to the previous study (14). The mRNA expression levels were normalized to the cycle threshold value of the housekeeping gene GAPDH (forward: 5’-ccactcctccacctttgac, reverse: 5’-accctgttgctgtagcca). All reagents and kits were purchased from Thermo Fisher Scientific.

The ChIP assay was performed as described in our previously published study (16). Equal amounts of DNA from each sample were used to perform qRT–PCR to quantitate the amount of DNA using primers designed for the IL33 promoter regions (forward: 5’-cagatctggagcagctgttc, reverse: 5’-aggccgtggtcactcatatt) according to the prediction from PROMO version 3.0.2 software. The relative amounts of the amplified products were normalized to the total input DNA amount in the samples.

The animal experiments followed the guidelines of the Animal Center of Kaohsiung Medical University and was approved by the Kaohsiung Medical University–Institutional Animal Care and Use Committee (IACUC-106054). We generated club cell-specific AhR KO mice (Scgb1a1-Cre; AhR^flox/flox ) as previously described (17). All mice, including wild-type (Scgb1a1-Cre) and Scgb1a1-Cre; AhR^flox/flox , were maintained in a specific pathogen-free facility, randomly divided into four groups, and orally fed 0.01% DMSO group in corn oil, 25μg DEHP/kg body weight (DEHP group) in corn oil, 20 mg/kg body weight kaempferol (K group) or both DEHP and kaempferol (DEHP+K group) in corn oil (Sigma–Aldrich) every day. The mice were sensitized with 100 µg OVA by intraperitoneal injection on day 4. The mice were challenged with 3% aerosolized OVA over three successive days (days 15–17). The mice were sacrificed on day 18 for examinations of cytokine production and histology. The experimental concentration of DEHP was referred to DEHP daily intake in children in the indoor environment (18). We choose a dosage of kaempferol that could suppress allergic airway inflammation as previously described (19). The flow-chart of the treatment in experimental animal showed in Supplementary Figure 1 .

Bronchoalveolar lavage fluid (BALF) procedures was generated as described previously with a minor modification (17). Tracheae was cannulated and the lung was lavaged 2 times with 1 mL PBS (contain 1% FCS). Each fluid was centrifuged and the supernatant was rapidly frozen at −80°C. The cells in BALF were stained with CD11c-PE/Cy7 (N418), MHC class II-FITC (M5/114.15.2), anti-B220 (RA3–6B2; eBioscience, Thermo Fisher Scientific), CCR3-PE (83,101; R&D Systems, Minneapolis, USA), and CD3-APC (145–2C11; BD Biosciences, San Diego, USA) and antibodies and analyzed by flow cytometry (LSR II; BD Biosciences) (17). The gating strategy of immune cells in BALF was showed in Supplementary Figure 2 .

Lung samples were fixed in formaldehyde, embedded in paraffin, cut into 3 μm thick sections, stained with hematoxylin and eosin for general inflammation analyses, and observed by a light microscope (Axio Imager M1; Carl Zeiss, Oberkochen, Germany) (17).

Concentrations of IL-4 and IL-33 in BALF or supernatant were determined by commercially ELISA systems using the protocol recommended by the manufacturer (R&D Systems).

Continuous variables were presented as the mean ± standard deviation (SD) were tabulated to describe the distribution. Pearson’s chi-square tests were used to evaluate the significance of differences in categorical variables between non-allergic groups and allergic groups. Independent sample t-test was used to examine the differences in total IgE or cord blood IgE between non-allergic groups and allergic groups; spearman correlation was used to examine the correlation between maternal urinary metabolite levels and cord blood cytokine levels in allergic groups. Statistical analysis was performed using SPSS (Version 20, IBM Company, Armonk, NY, USA) or GraphPad Prism (Version 5, Los Angeles, CA, USA). The bar graphs display the means ± standard deviation (SD). Statistical analysis was performed using SPSS (Version 20, IBM Company, Armonk, NY, USA) or GraphPad Prism (Version 5, Los Angeles, CA, USA). The Mann–Whitney U test was used to determine differences between control and experimental group. The one-way ANOVA with Bonferroni post-test was used to analysis the results of the animal data. A p value < 0.05 was considered to indicate a significant difference.

IL-33 has recently been strongly recognized for its role in asthma and allergic diseases, the earliest manifestations of childhood allergy (5). We analyzed the cord blood (CB) IL-33 levels to examine the susceptibility of the newborn immune system to the potential adverse effects of in utero exposure to phthalates. The CB IL-33 levels in the allergic group were significantly higher than those in the non-allergic group ( Figure 1A ). We further investigated the levels of maternal urinary DEHP metabolites concentrations on the phenotype of allergic diseases. DEHP is rapidly metabolized to its monoester, MEHP, which is further metabolized by various hydroxylation and oxidation reactions to secondary metabolites, e.g., mono(2-ethyl-5-oxohexyl) phthalate (MEOHP), mono(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), mono(2-ethyl-5-carboxypentyl) phthalate (MECPP), and mono(2-carboxymethylhexyl) phthalate (MCMHP) (20). In particular, these metabolites appear suitable for assessing human DEHP exposure (biological monitoring) because they are not subject to external contamination. The concentrations of five common urinary DEHP metabolites in pregnant women were significantly higher in the allergic group than in the non-allergic group ( Figures 1B-F ). We next analyzed whether the CB IL-33 level was associated with maternal urinary DEHP metabolites level in the allergic group. The results showed a significantly positive correlation between MEHP and CB IL-33 ( Figure 1G ), but such a relationship was not shown for the other secondary metabolites ( Figures 1H-K ). We also observed whether the difference of CB IL-33 levels in different clusters including maternal allergic history, pets or not, second-hand smoking exposure during pregnancy, and the severity of allergic rhinitis in child. In the maternal allergic history cluster, the results showed the level of CB IL-33 was no difference in mother with or without allergic history. In the other clusters, the level of CB IL-33 also observed no difference between each two groups ( Supplementary Table 1 ). We further used the geometric mean (geo mean) to divide samples into low or high groups (21) based on maternal urinary MEHP (geo mean: 1.944 µg/g creatinine) and CB IL-33 levels (geo mean: 39.34 pg/mL). We distinguished the children in allergic groups into four groups, group 1 (elevated CB IL-33 and diminished maternal urinary MEHP), group 2 (diminished CB IL-33 and diminished maternal urinary MEHP), group 3 (diminished CB IL-33 and heightened maternal urinary MEHP), and group 4 (elevated CB IL-33 and heightened maternal urinary MEHP). We want to investigate the potential dissimilarities in phenotypic expression among these four groups. In Supplementary Table 2 , the maternal age, exposure to second-hand smoking during pregnancy, maternal allergic history, children’s age, gender, delivery methods, exposure to second-hand smoking, and raised pets showed no difference in these four groups. The prevalence of asthma, allergic rhinitis, and atopic dermatitis also showed no difference in these four groups. Otherwise, the secondary maternal urinary DEHP metabolites (MEOHP, MEHHP, MECPP, and MCMHP) were also elevated in group 3 and group 4. However, the CB total IgE and total IgE were also no different between these four groups. Even the total IgE in group 1 and 4 seemed higher than group 2 or group 3, but it was no statistical significance.

Since the CB IL-33 levels were significantly higher in the allergic group and high MEHP group, to examine whether DEHP and its important metabolite MEHP affect IL-33 production in human lung and bronchial epithelial cells. As shown in Figure 2 , DEHP or MEHP induced IL33 mRNA expression ( Figures 2A, B ) in a dose-dependent manner in A549 cells. In HBE cells, IL33 mRNA expression was also increased in a dose-dependent manner after DEHP or MEHP treatment ( Figures 2C, D ). We further used HDM (Dermatophagoides pteronyssinus) extract as a common allergen to evaluate the effect of DEHP or MEHP on IL-33 production. We observed that DEHP or MEHP treatment enhanced IL-33 mRNA expression ( Figure 2E ) and cytokine production ( Figure 2F ) in A549 cells. However, DEHP or MEHP alone did not induce IL-33 protein production (data not shown).

We further added an AhR antagonist to investigate the suppressive effect on DEHP- or MEHP-induced IL-33 production. Our results showed that the expression of IL33 mRNA induced by DEHP or MEHP stimulation were reversed by the AhR antagonist ( Figures 3A, B ). The increase in HDM-induced IL-33 production induced by DEHP or MEHP treatment was also reduced by the AhR antagonist ( Figures 3C, D ). These results suggested that the regulatory effects of phthalates on IL-33 might be mediated through nuclear AhR. Flavonoids have attracted great attention for their anticancer and anti-inflammatory activities (22). Therefore, we hypothesized that flavonoids might suppress the expression of IL-33 through nuclear AhR. The results showed that DEHP and MEHP-induced IL33 mRNA expression was suppressed by kaempferol and apigenin ( Figures 3E-H ). The increase in HDM-induced IL-33 production induced by DEHP or MEHP treatment was reduced by kaempferol and apigenin ( Figures 3I, J ). We also chose the AhR agonist, 2,3,7,8-tetrachlorodibenzodioxin (TCDD), which dose referred to the previous study (23), to observe whether the AhR was involved in MEHP-induced IL-33 expression. In Supplementary Figure 3 , the expression of IL-33 mRNA was significantly increased by MEHP or TCDD treatment and suppressed by pretreatment with AhR antagonist or kaempferol. These results implied that the suppressive effect of flavonoids on IL-33 expression might be mediated through AhR.

Histone modifications are a component of epigenetic regulation and an important modulator of gene expression (24). We used AA and MTA to evaluate whether histone modifications were involved in the DEHP- or MEHP-induced increases in the expression of the IL33 mRNAs. The increased level of IL33 mRNA induced by DEHP or MEHP was decreased by AA ( Figures 4A, B ). Moreover, the DEHP- or MEHP-induced increases in IL33 mRNA expression were also suppressed by MTA ( Figures 4C, D ). Histone H3 and H4 acetylation in the IL33 promoter region were upregulated by DEHP or MEHP stimulation, especially at high doses ( Figures 4E, F ). Furthermore, histone H3K4, H3K36, and H3K79 trimethylation in the IL33 promoter region were also upregulated ( Figures 4G, H ). We also observed whether the AhR was involved in histone modification by MEHP stimulation. In Supplementary Figure 4A , the upregulated acetylation level of histone H3 and H4 by MEHP stimulation were reduced by AhR treatment. The MEHP-regulated tri-methyl H3K4, H3K36, and H3K79 were also downregulated by AhR treatment. These results show that DEHP- or MEHP-induced changes in the expression of the IL33 mRNAs are regulated by histone H3 and H4 acetylation and histone H3K4, H3K36, and H3K79 trimethylation via AhR.

In Scgb1a1-Cre mice, the number of total cells in bronchoalveolar lavage fluid (BALF) in the DEHP injection group was significantly increased compared with that in the dimethyl sulfoxide (DMSO) group and was significantly reduced after co-feeding of kaempferol. However, the suppressed effect of kaempferol did not shown in Scgb1a1-Cre; AhR^flox/flox mice ( Figure 5A ). The counts of neutrophils, eosinophils, and T cells in BALF showed similar patterns both in Scgb1a1-Cre and Scgb1a1-Cre; AhR^flox/flox mice ( Figures 5B-D ). We also measured the production of the Th2 cytokines IL-4 and IL-33 in BALF. OVA-induced IL-4 and IL-33 production was significantly enhanced by DEHP and suppressed by kaempferol in Scgb1a1-Cre mice ( Figures 5E, F , left side). We further observed the cytokines IL-33 and IL-4 in Scgb1a1-Cre; AhR^flox/flox to assess the role of AhR in the DEHP response. In Scgb1a1-Cre; AhR^flox/flox mice, the IL-33 and IL-4 concentrations were not different between the DMSO and DEHP groups, and IL-33 was significantly decreased by kaempferol ( Figures 5E, F , right side). Histological staining showed that exposure to DEHP further enhanced lung inflammation in Scgb1a1-Cre mice but not in Scgb1a1-Cre; AhR^flox/flox mice, and was alleviated by kaempferol ( Figure 5G ). The results suggested that DEHP enhances lung inflammation and the production of the IL-33 and IL-4 through AhR.