Diet-induced obesity worsens allergen-induced type 2/type 17 inflammation in airways by enhancing DUOX1 activation

Abstract

More than 50% of people with asthma in the United States are obese, and obesity often worsens symptoms of allergic asthma and impairs response to treatment. Based on previously established roles of the epithelial NADPH oxidase DUOX1 in allergic airway inflammation, we addressed the potential involvement of DUOX1 in altered allergic inflammation in the context of obesity. Intranasal house dust mite (HDM) allergen challenge of subjects with allergic asthma induced rapid secretion of IL-33, then IL-13, into the nasal lumen, responses that were significantly enhanced in obese asthmatic subjects (BMI >30). Induction of diet-induced obesity (DIO) in mice by high-fat diet (HFD) feeding similarly enhanced acute airway responses to intranasal HDM challenge, particularly with respect to secretion of IL-33 and type 2/type 3 cytokines, and this was associated with enhanced epithelial DUOX1 expression and was avoided in DUOX1-deficient mice. DIO also enhanced DUOX1-dependent features of chronic HDM-induced allergic inflammation. Although DUOX1 did not affect overall weight gain by HFD feeding, it contributed to glucose intolerance, suggesting a role in glucose metabolism. However, glucose intolerance induced by short-term HFD feeding, in the absence of adiposity, was not sufficient to alter HDM-induced acute airway responses. DIO was associated with enhanced presence of the adipokine leptin in the airways, and leptin enhanced DUOX1-dependent IL-13 and mucin production in airway epithelial cells. In conclusion, augmented inflammatory airway responses to HDM in obesity are associated with increases in airway epithelial DUOX1, and by increased airway epithelial leptin signaling.

INTRODUCTION

Obesity is a major risk factor for a number of health problems, such as cardiovascular disease and type 2 diabetes, and is also associated strongly with an increased risk of asthma development and asthma severity. The incidences of obesity and asthma have increased dramatically over the past 30 years in the United States and worldwide, and obesity is more common in people with asthma in the United States compared with the general population (1–3). Adults with obesity often develop nonatopic asthma. However, increased risk of atopy and T_H2 responses have also been observed in patients with obesity, and obesity also contributes to asthma exacerbation in children in whom the disease is mostly allergic (4, 5).

Despite intense research over the past decades, the mechanisms by which obesity leads to worsened asthma symptoms still remain poorly understood, and likely include mechanical effects linked to adipose tissue accumulation, obesity-associated inflammation, metabolic dysfunction, and increased oxidative stress that is associated with obesity. Studies in mice have indicated that diet-induced obesity can exacerbate allergic airway inflammation by enhanced Th2 and Th17 inflammatory responses due to increased lung populations of type 2 and type 3 innate lymphoid cell (ILC) populations and production of IL-33 and IL-1B as main activators of these cell populations (6, 7). Moreover, T_H2 and ILC2 responses may be enhanced by alterations in adipokine levels due to obesity (8–10), and may also be altered due to insulin resistance (11–13).

One of the main features of metabolic alterations during obesity that has been linked with increased disease pathology is oxidative stress (14, 15), which may also contribute to exacerbated features of allergic asthma (16, 17). Such oxidative stress is thought to originate largely from mitochondrial dysfunction, caused by excessive nutrient intake leading to high concentrations of free fatty acids and hyperglycemia, which leads to decreased TCA cycle metabolism and increased glycolysis (18, 19), reduced mitochondrial biogenesis, and impaired mitochondrial fusion/fission dynamics (20–24), which collectively results in increased mitochondrial production of reactive oxygen species (ROS) (15, 25, 26). The NOX family of NADPH oxidases represents additional major sources of ROS and may play important roles in inflammation and asthma pathology (27, 28). Indeed, our recent studies have highlighted an important role for the NADPH oxidase (NOX) enzyme DUOX1 in innate airway responses to airborne allergens and in type 2 inflammation during allergic asthma (29–31). However, although altered activation of several NOX enzymes has been associated with metabolic dysfunction during obesity (32), the relationship of obesity with DUOX1 has not been addressed to date.

The present studies were conducted to assess the impact of diet-induced obesity (DIO) on allergic airway inflammation and its potential association with altered DUOX1 expression or function. We demonstrate that DIO results in enhanced DUOX1 expression within the airway epithelium, which contributes to enhanced type 2 and type 3 inflammatory responses during house dust mite (HDM)-induced allergic airway inflammation. DUOX1 did not affect overall weight gain during DIO, but contributed to alterations in glycemic control, as a potential additional contributing factor to enhanced features of allergic inflammation associated with obesity. Finally, our findings also reveal a contribution of DUOX1 to leptin-induced proinflammatory responses and mucus production within the airway epithelium.

METHODS

Human Subjects

Human nasal epithelial (HNE) cells collected from participants with allergic asthma were cultured for RNA extraction and analysis of proteins cysteine oxidation (sulfenylation). Details of the study have been previously reported (31); we included participants with allergic rhinitis and asthma and normal healthy controls without rhinitis. A separate cohort of lean (body mass index (BMI) 18.5–24.9 kg/m²) and obese (BMI ≥ 30 kg/m²) individuals with allergic rhinitis and asthma, and with positive skin prick test to Dermatophagoides pteronyssinus, were enrolled for intranasal challenge studies with HDM (∼200 AU D. pteronyssinus antigen extract, in 50% glycerol, GB70A01, Lot No. 322998, Greer Laboratories, NC) administered intranasally under IND No. 15214. Nasal lavage with 5 mL of sterile saline was collected before, and 15 min or 24 h after the challenge, for analysis of IL-33 and IL-13 by ELISA (R&D Systems) and for analysis of ATP and H₂O₂ as described previously (31). The identity of H₂O₂ in nasal lavages was verified by its removal with catalase (Worthington). All studies were approved by the University of Vermont Institutional Review Board (Protocol Numbers 15-067 and 13-167), and written informed consent was obtained from all participants.

Mouse Studies

Duox1-deficient (Duox1^−/−) mice, originally generated on the C57BL/6J background (33), were backcrossed for at least five generations onto the C57BL/6NJ background (Stock No. 005304; Jackson Laboratories, Bar Harbor, ME), to avoid potential alterations in susceptibility to diet-induced metabolic disease in the 6J strain due to genetic variances including loss of mitochondrial nicotinamide nucleotide transhydrogenase (NNT) (34). Some studies were also performed with Scgb1a1^Cre-ERTDuox1^F/F mice that were generated as described previously (35), and in which DUOX1 is conditionally deleted from airway epithelial cells upon intraperitoneal injection of tamoxifen (Sigma-Aldrich; No. 10540-29-1). Both wild-type and DUOX1-deficient mouse strains were bred in-house under pathogen-free conditions with constant room temperature and humidity and a 12-h light cycle. Starting at 6–8 wk of age, mice were randomly assigned to feeding protocols with either high-fat diet (HFD, with 60% kcal from fat; Research Diets D12492) or low-fat diet (LFD, with 10% kcal from fat; Research Diets D12450B) for up to 14 wk before experimentation. Weight gain was monitored weekly throughout the feeding protocol, and in some cases, blood was collected for analysis of leptin, insulin, or glucose. Some mice were also subjected to an intraperitoneal glucose tolerance test as described later. Both male and female mice were used and were randomly distributed among experimental groups.

Following either HFD or LFD feeding, mice were subjected to a single airway challenge with 50 μg of house dust mite (HDM) extract (D. pteronyssinus; Greer Laboratories; Lot No. 269026 and 305470; 32–64 endotoxin U/mg; 6–9 µg Der p1/mg) in 50 μL of PBS, which was administered oropharyngeally under isoflurane anesthesia. Control mice received 50 μL of PBS vehicle control. One (1) h after acute HDM challenge, BAL fluids, lung tissues, and visceral adipose tissues (VAT) were collected for various analyses described in the next sections. In separate studies, mice were subjected to HDM-induced allergic airway inflammation, by sensitizing them by airway HDM instillation (50 µg) on days 1 and 8, followed by daily airway challenge with 50 μg HDM on days 15–19. Bronchoalveolar lavage (BAL) fluids and lung tissues were collected three days after the final HDM challenge (day 22) for analysis. All studies involving mice were approved by the Institutional Animal Care and Use Committee (Protocol No. PROTO202000078).

Analysis of Lung Inflammation

BAL fluids were collected by three successive instillations of 500 µL of PBS into the lungs and pooled BAL fluids were centrifuged at 150 g after which BAL cells were collected for cell differentials and supernatants were collected for analysis of cytokines and other secreted factors. BAL levels of IL-33, IL-25 (IL-17E), IL-1α, IL-1β, IL-5, IL-13, and IL-17A were assessed using DuoSet ELISA Kits (R&D Systems) according to the manufacturer’s protocol. Pelleted BAL cells were resuspended in 1 mL of PBS and counted using a hemocytometer. The remaining cells were loaded into an EZ Cytofunnel (Thermo Scientific) and centrifuged at 600 rpm for 10 min, to generate cytospin slides that were fixed and stained using the Hema 3 kit (Thermo Scientific). Cell differentials were counted based on at least 200 cells per slide.

Analysis of BAL and Serum Insulin and Leptin

Nonfasting blood (up to 300 µL) was collected at the time of animal harvest from the posterior vena cava, and serum and BAL fluids were evaluated for insulin or leptin by ELISA (Crystal Chem Ultra-Sensitive Mouse Insulin ELISA; Cat. No. 90080, and R&D Systems Quantikine Elisa Mouse/Rat Leptin Cat. No. M0B00B).

RT-PCR Analysis

Total RNA was extracted from mouse inferior lobes or from VAT using GeneJet RNA purification kits (Invitrogen) according to the manufacturer’s protocol. cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad) and/or Applied Biosystems High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) according to the manufacturer’s protocol. Quantitative RT-PCR was performed using iQ Sybr Green supermix (Bio-Rad), 0.5 µL of cDNA, and 0.5 µM primer mix (Supplemental Table S1) as previously described, using GADPH as housekeeping gene (36).

Immunohistochemistry Analysis of DUOX1

Left lung lobes were fixed in PFA and paraffin embedded, and 5-μm sections were stained with α-DUOX1 (Santa Cruz, SC48858; 1:100) and utilized a biotin-conjugated secondary antibody (Dako, E0466), and visualized using Vectastain Peroxidase ABC/AP Kit (Vector Laboratories, AK-5000) and Enzyme Substrate (Vector Blue SK5300; Vector Laboratories) with Nuclear Fast Red counterstaining (H3403).

Flow Cytometry Analysis

Single-cell suspensions were generated from lung by enzymatic digestion with the Lung Dissociation Kit (Miltenyi Biotec 130-095-927) and a GentleMACS Dissociator (Miltenyi Biotec) following the manufacturer’s protocol. Visceral adipose tissues (VAT) were digested to collect stromal vascular fraction (SVF), containing preadipocytes and myeloid/lymphoid cells, as described previously (37). One million cells of lung or SVF cell suspensions were incubated in Live/Dead Fix Blue (Thermo Fisher L23105) for 20 min at 4°C and then washed and incubated with Fc Block (anti-mouse CD16/32) (BD Pharminogen 553142) for 10 min at room temperature and appropriate antibodies were added for 30 min at 4°C [Alexa700-Lin- cocktail (Biolegend 133313), eFluor450-CD90.2 (eBioscience 48-0902-80), APC-CD25 (Biolegend 101909), PE-Cy7-T1/ST2 (Biolegend 145315), PE-CD127+ (Biolegend 135009), and PerCP/Cy5.5-Nkp46 (Biolegend 137609)]. Cells were then washed twice, fixed with 1% paraformaldehyde, and analyzed using a BD LSRII flow cytometer and FlowJo v10.7 software. ILC2/3 populations were analyzed as Lin⁻CD25^highCD90^high populations and distinguished based on ST2^high (ILC2) or ST2^lowNkp46^highCD127^high (ILC3). For characterization of VAT macrophage subtypes, one million cells per sample were incubated in Live/Dead Fix Blue (Thermo Fisher L23105), washed and incubated with Fc Block (anti-mouse CD16/32) (BD Pharminogen 553142), and incubated with the following antibody cocktail: Pacific Orange anti-CD45 (Invitrogen MCD4530), PE anti-F4/80 (BioLegend 123109), APC-eFluor780 anti-CD11b (Invitrogen 47-0112-82), PE-Cy7 anti CD11c (BioLegend 117317), and AlexaFluor647 anti-CD301 (BioLegend 145603). Cells were then washed, fixed, and analyzed on a BD LSRII flow cytometer. For characterization of M1- and M2-polarized macrophage populations, CD45⁺F4/80^highCD11b^high populations were distinguished based on surface expression of CD11c (M1 marker) or CD301 (M2 marker) (38).

Analysis of Protein Cysteine Oxidation and Epidermal Growth Factor Receptor Activation

For analysis of protein cysteine oxidation to sulfenic acids (-S-OH), cells or lung tissues were homogenized in deoxygenated Western solubilization buffer containing 1 mmol/L of the sulfenic acid probe DCP-Bio1 (Kerafast, Boston, Mass), 200 U/mL of catalase (Worthington, Lakewood, NJ), and 10 mmol/L of N-ethylmaleimide (Sigma) and incubated for 1 h at 4°C, as described previously (30). Derivatized homogenates were subsequently analyzed by SDS-PAGE and Western blotting with streptavidin-HRP. To isolate biotin-tagged proteins, excess biotinylating reagent was removed by six successive washes with 20 mmol/L Tris-HCl (pH 7.4) on Amicon Ultra-0.5 Centrifugal Filter Devices (Millipore, Temecula, CA), and samples were mixed overnight with NeutrAvidin agarose beads (50 µL of 50/50 slurry, Pierce) and washed successively with 1% SDS, 4 mol/L of urea, 1 mol/L of NaCl, and 100 mmol/L of ammonium bicarbonate to remove nonspecifically bound proteins (31). Biotinylated proteins were eluted using 2× reducing sample buffer containing β-mercaptoethanol and brief heating (10 min at 90°C), and eluted proteins were analyzed by SDS-PAGE and Western blotting for epidermal growth factor receptor (EGFR) (Cell Signaling; C74B9; 1:1,000). Whole tissue homogenates or cell lysates were analyzed similarly for EGFR as input controls. EGFR activation was assessed by Western blotting with phosphorylated EGFR (pEGFR Y1068; Cell Signaling; D7A5, 1:1,000). In each case, blots were visualized using enhanced chemiluminescence (Pierce).

Glucose Tolerance Test

LFD- or HFD-fed mice were fasted for 16 h before performing an intraperitoneal glucose tolerance test (IPGTT). After an initial blood sample, mice were injected with d-glucose (1.5 mg/g body weight), and glucose concentrations in blood samples taken from the tail vein at 15, 30, 60, 120, and 240 min were measured using a OneTouch Ultra2 blood glucose meter. Glucose concentrations were plotted against time, and area under the curve (AUC) was calculated using GraphPad Prism (v.8.2.1).

Studies Using Cultured Airway Basal Cells

Mouse airway basal cells were cultured from tracheas collected from HFD- or LFD-fed mice, by previously described protocols (31), and used at passage 1–2 for in vitro experimentation. For experimentation, cultured airway basal cells were seeded to confluence in 24-well culture dishes, cultured overnight in EGFR-free media, and stimulated with either HDM or leptin for up to 24 h. Conditioned media or cell lysates were collected for ELISA, RNA extraction, and RT-PCR analysis, or analysis of EGFR sulfenylation or activation.

Statistical Analysis

Experiments in mice were performed twice for each diet group and the data were pooled and analyzed by Student’s t test or by one- or two-way analysis of variance (ANOVA), and Tukey’s test for multiple comparisons, using GraphPad Prism (v8.2.1). All quantitative results are presented in dot-plots and highlighted with mean values ± SE. Differences were considered statistically significant when P values were <0.05.

RESULTS

Innate Nasal Epithelial Type 2 Responses to Acute HDM Are Enhanced in Obese Subjects

We recently evaluated DUOX1 mRNA and protein expression in nasal epithelial cells (NCEs) cultured from epithelial brushes from 10 subjects with allergic asthma (IgE levels > 100) and 10 nonasthma controls (31). Using available BMI data for these subjects, we correlated DUOX1 mRNA expression in these NCEs with BMI of these subjects, which revealed a significant positive association (Fig. 1A). DUOX1 generates H₂O₂ that can oxidize protein cysteine thiols to sulfenic acids. Assessment of protein sulfenic acid intermediates in NCE protein extracts using derivatization with DCP-Bio1 revealed elevated sulfenic acid levels in NCEs from lean asthmatic subjects (BMI < 25) compared with lean nonasthmatic controls, with further increases apparent in NCEs from allergic asthmatics that were also obese (BMI > 30) (Fig. 1B).

Figure 1.

Obesity is associated with increased nasal DUOX1 and innate type 2 cytokines response to house dust mite (HDM) challenge. A: correlation of DUOX1 mRNA expression in nasal epithelial (HNE) cells obtained from nonasthmatic participants (black symbols) and patients with allergic asthma (red symbols) with BMI. B: analysis of overall protein sulfenylation in HNE cells from nonasthmatic controls and from lean or obese patients with allergic asthma, using derivatization with the sulfenic acid probe DCP-Bio1 and SDS-PAGE and streptavidin blot. C–F: analysis of nasal lavage fluids collected before or after intranasal HDM challenge of lean and obese participants with allergic asthma for IL-33 (C), IL-13 (D), ATP (E), or H₂O₂ (F). *P < 0.05; **P < 0.01; ****P < 0.0001; 2-way ANOVA; n = 6 subjects/group.

Based on these observations, and our prior studies showing that DUOX1-derived H₂O₂ promotes secretion of the alarmin IL-33 (31), we performed intranasal HDM challenge in a subset of lean and obese people with house dust mite allergic asthma, and collected nasal lavage fluids to monitor acute nasal responses. As expected (31, 39), HDM challenge resulted in acute increased and transient production of IL-33 into the nasal secretions, and more delayed and sustained secretion of IL-13, responses that are believed to be linked to activation of DUOX1 within the airway epithelium (31) (Fig. 1, C and D). Strikingly, both responses were significantly enhanced in obese subjects compared with lean subjects, consistent with relatively higher levels of DUOX1, and in line with previous in vitro studies of nasal epithelial cells showing that increased DUOX1 expression (e.g., in the context of allergic asthma) contributes to increased IL-33 secretion upon HDM challenge (31). To assess apparent activation of DUOX1 upon in vivo nasal HDM challenge, we measured acute production of extracellular ATP, a known damage signal that activates DUOX1, as well as H₂O₂, a product of DUOX1 activation. Indeed, although both ATP and H₂O₂ were undetectable at baseline (<10 nM), HDM induced transient extracellular ATP production, which was similar in both subject groups (Fig. 1E), and also induced H₂O₂ production, which tended to be higher in obese compared with lean subjects, although this was not statistically significant (Fig. 1F). Neither ATP nor H₂O₂ was detectable at 24 h after HDM challenge. Overall, these findings are consistent with previous studies documenting enhanced Th2 cytokine responses in obesity-associated asthma (4, 5), and indicate increased activation of type 2 cytokine responses by intranasal allergen challenge in obese subjects, which appears to be related to increased expression or activation of DUOX1 within the nasal epithelium.

Diet-Induced Obesity Exacerbates DUOX1-Dependent Proinflammatory Cytokine Responses to Acute Airway HDM Challenge in Mice

We performed comparative studies of acute intranasal HDM challenge in mice subjected to diet-induced obesity (DIO), and assessed the contribution of DUOX1 by performing similar studies with DUOX1-deficient mice (Fig. 2A). As expected, mice fed HFD over 14 wk gained more weight compared with mice on LFD (Fig. 2B), and weight gain was in both cases somewhat lower in female mice (Fig. 2C). Also, consistent with earlier findings showing that body fat distribution differs between males and females (40), female mice displayed more VAT weight gain upon DIO compared with males (Fig. 2D). Importantly, overall weight gain and VAT weight gain by HFD feeding were not altered in DUOX1-deficient mice (Fig. 2, B–D), which also didn’t significantly affect DIO-induced changes in VAT macrophage polarization (Supplemental Fig. S1). Consistent with data from human nasal epithelial cells (Fig. 1), DIO in mice resulted in increased DUOX1 mRNA expression in the lung (Fig. 2E), and increased DUOX1 protein expression within the airway epithelium (Fig. 2F). It is worth noting that such DUOX1 mRNA increases are not unique to DIO, and DUOX1 mRNA was also enhanced in lung tissues from obese Lepr^db/db mice lacking the long form of the leptin receptor (Jackson Laboratory), compared with lung tissues from normal Lepr^+/+ C57BL/6J mice (Supplemental Fig. S2A). Also, DIO resulted in altered redox homeostasis in lung tissues, as illustrated by increased overall cysteine oxidation to sulfenic acids (Supplemental Fig. S2B).

Figure 2.

Diet-induced obesity (DIO) is associated with increased DUOX1 expression in the airway. A: schematic of animal feeding with low-fat or high-fat diets (LFD, HFD) and subsequent exposure to house dust mite (HDM) allergen. B: average overall weekly weights of male wild-type (WT) or Duox1^−/− mice fed LFD or HFD over the course of 14 weeks (n = 4). C: weight gain of male (M) or female (F) WT or Duox1^−/− mice after 14 wk of LFD or HFD feeding. D: visceral adipose tissues (VAT) weight in same animal groups. E: effect of HFD on Duox1 mRNA expression in lung tissues. F: effect of HFD of DUOX1 protein expression by immunohistochemistry. Data from two representative animals are shown for both groups. **P < 0.01; ***P < 0.001; ****P < 0.0001; 2-way ANOVA; n = 8–10/group.

Previous studies suggest that DIO in mice may affect innate airway responses to allergen by increases in ILC2 and ILC3 cell populations within the lung (6, 7), but the impact of DIO on innate epithelial responses to airway allergen challenge is not known. Consistent with results from the nasal challenge studies, acute airway HDM challenge resulted in greater secretion of the epithelial alarmin IL-33 in HFD-fed mice compared with nonobese (LFD-fed) mice (Fig. 3A), which was not accompanied by increased lung IL-33 mRNA expression (Fig. 3F), thus suggesting a specific effect on IL-33 secretion, presumably from the epithelium (31). Similarly, DIO also appeared to enhance other innate alarmin responses to HDM, including IL-25 and IL-1α (Fig. 3, B and C), although differences were not statistically significant, and enhanced HDM-mediated induction of Il5 and Il13 mRNA levels in lung tissues, being significant in the latter case (Fig. 3, G and H). In each of these cases, these responses were almost fully suppressed in DUOX1-deficient mice, in which DIO-dependent increases were not observed. Although previous studies indicated some sex differences with respect to DUOX1 regulation (41) or type 2 inflammation (42), our studies were not sufficiently powered to reveal statistically significant differences between male and female mice with respect HDM-induced responses or its alterations by DIO (not shown).

Figure 3.

Diet-induced obesity (DIO) results in enhanced airway responses to house dust mite (HDM) allergen challenge. Mice were subjected to low-fat diet (LFD) or high-fat diet (HFD) for 14 wk and then challenged with HDM for 1 h. Bronchoalveolar lavage (BAL) fluids were analyzed for IL-33 (A), IL-25 (B), IL-1α (C), IL-17A (D), or IL-1β (E) by ELISA, and lung tissues were analyzed for mRNA expression of Il33 (F), Il5 (G), Il13 (H), Il17a (I), or Il22 (J) by RT-PCR. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; 2-way ANOVA; n = 4–7/group.

To address the potential involvement of Th17 cells or type 3 ILCs in enhanced innate airway responses due to DIO, we evaluated the production of IL-1β, IL-17A, and IL-22, by BAL ELISA or by RT-PCR analysis of lung tissues. Indeed, HDM challenge slightly increased BAL levels of IL-17A and markedly enhanced IL-1β, which tended to be increased in the context of DIO (Fig. 3, D and E). Both responses were also attenuated in Duox1^−/− mice. Similarly, HDM challenge enhanced lung tissue Il17a mRNA (Fig. 3I), which was enhanced further in the context of DIO and attenuated in Duox1^−/− mice. Finally, lung tissue expression of Il22, a cytokine primarily produced by Th17 cells or type 3 ILCs (43), was enhanced by DIO even under basal conditions and was further increased upon HDM challenge, but these responses were not altered in DUOX1-deficient animals (Fig. 3J).

To assess whether DIO-mediated changes in DUOX1 and innate responses indeed occur at the level of the respiratory epithelium, we cultured airway basal cells from tracheas of LFD- or HFD-fed mice for in vitro HDM challenge studies. Indeed, cultured airway basal cells from HDF-fed mice displayed increased relative DUOX1 mRNA expression, compared with airway basal cells obtained from LFD-fed mice (Fig. 4A), and showed augmented response to in vitro HDM stimulation with respect to IL-33 secretion (Fig. 4B). Based on previous studies implicating DUOX1-dependent redox-based activation of EGFR within the airway epithelium (31), we examined HDM-induced phosphorylation and sulfenylation of EGFR in airway basal cells, which revealed that both were enhanced in response to HDM or to ATP (a damage signal released by HDM), and this occurred to a greater extent in airway basal cells from HFD-fed mice (Fig. 4C). Similarly, EGFR sulfenylation was also enhanced in mouse lung tissues upon acute in vivo HDM airway challenge, which was more substantial in the context of DIO, and was almost completely prevented in Duox1^−/− mice (Fig. 4D).

Figure 4.

Diet-induced obesity (DIO) leads to enhanced DUOX1 and house dust mite (HDM)-responses in airway basal epithelial cells in vitro. Cultured airway basal epithelial cells isolated from low-fat diet (LFD)- or high-fat diet (HFD)-fed mice were analyzed for Duox1 mRNA (A), or IL-33 secretion into the media after 1-h HDM stimulation (B). **P < 0.01; ****P < 0.0001; t test or 2-way ANOVA; n = 6 replicates from two independent experiments. C: analysis of EGFR phosphorylation or sulfenylation in lysates of basal airway epithelial cells by Western blot. D: analysis of epidermal growth factor receptor (EGFR) sulfenylation in lung tissues from LFD- or HFD-fed mice challenged with HDM (1 h). Blots representative of two independent experiments are shown.

Collectively, these findings indicate that DIO leads to enhanced type 2 and type 3 responses to acute airway HDM challenge, which is primarily linked to increased epithelial DUOX1 expression and early DUOX1-dependent production of epithelial cytokines such as IL-33 and IL-1β, which in turn activate Th2 or Th17 responses. Although some of our observations (e.g., increased Il22 mRNA in unchallenged DIO mice) may be related to increases in lung ILC2/3 cells populations in the context of DIO (6, 7), efforts to quantify lung ILC2/3 cell populations by flow cytometry did not reveal significant differences in the context of DIO or DUOX1-deficiency (Supplemental Fig. S3).

Impact of DIO HDM-Induced Allergic Inflammation and Remodeling

We next assessed the impact of DIO and DUOX1 on features of allergic airway inflammation and remodeling in response to HDM sensitization and repeated challenge over a 3-wk period. As shown in Fig. 5A, total BAL cell numbers were increased upon repeated HDM challenge but were not significantly different between LFD and HFD groups, or between wild-type (WT) and Duox1 knockout (KO) mice. BAL cell differentials showed increases in lymphocytes and eosinophils in HDM groups, as well as neutrophils. DIO tended to increase eosinophil recruitment by HDM, whereas neutrophil numbers tended to decrease, but neither was statistically significant (Supplemental Fig. S4A). Consistent with previous studies (30), DUOX1 deficiency did not significantly affect eosinophil recruitment, but attenuated neutrophil recruitment in both LFD and HFD groups. Production of type 2 inflammation during chronic HDM sensitization and challenge was generally enhanced after DIO, with greater increases in BAL IL-33, IL-5, and IL-13 (Fig. 5, B–D), and this was also reflected in increased lung tissue mRNA levels for these cytokines (significant only in case of IL-13; Supplemental Fig. S4B). Likewise, consistent with increased lung tissue expression of Duox1 by HDM (Supplemental Fig. S4B), lung tissues of HDM-treated mice also displayed increased protein cysteine sulfenylation (Supplemental Fig. S2B), and these increases were in both cases more dramatic in the context of DIO. Repeated HDM challenge also enhanced lung tissue markers of mucus metaplasia (Muc5ac, Clca1) or subepithelial fibrosis (Col1a1, Col3a1), which tended to be enhanced to a greater degree in HFD-fed mice, although this was not statistically significant. In both LFD and HFD groups, these HDM-induced increases in airway remodeling were significantly attenuated in DUOX1-KO mice (Fig. 5, E–H).

Figure 5.

Diet-induced obesity (DIO) affects house dust mite (HDM)-induced allergic airway inflammation and remodeling. Low-fat diet (LFD)- or high-fat diet (HFD)-fed mice were subjected to HDM sensitization and challenge to induce allergic airway inflammation. Bronchoalveolar lavage (BAL) fluids were analyzed for total cell counts (A), IL-33 (B), IL-5 (C), and IL-13 (D). Lung tissues were harvested for analysis of mRNA levels of Clca1 (E), Muc5ac (F), Col1a1 (G), and Col3a1 (H). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; 2-way ANOVA; n = 4–8 from two independent experiments.

Because DIO was associated with increased airway epithelial DUOX1, we also performed similar studies in which DUOX1 was conditionally ablated from the airway epithelium (35). Again, DIO resulted in increased airway type 2 cytokine production after chronic HDM stimulation, which was largely diminished after epithelial DUOX1 ablation (Supplemental Fig. S5). Overall, these data indicate that DIO enhances features of type 2 inflammation during HDM-induced allergic inflammation, particularly with respect to type 2 cytokine production, and that these changes are largely absent in the context of airway epithelial DUOX1 deficiency. Thus, while alterations in e.g., lung ILC profiles or systemic metabolic changes likely contribute to altered features of allergic inflammation in the context of DIO, increased epithelial expression and activation of DUOX1 represents an important factor in these outcomes.

Impact of DUOX1 on Glucose Homeostasis: Relevance for Type 2 Immune Responses?

Epidemiological studies suggest an association between allergic asthma and glycemic control and insulin resistance (11, 13, 44). Although we didn’t observe a significant effect of DUOX1 on HFD-induced weight gain, DIO resulted in increased nonfasting glucose, especially in male mice, and this was significantly attenuated in Duox1^−/− mice (Fig. 6A), suggesting some role for DUOX1 in altered glycemic control in the context of obesity. We also observed increases in serum insulin upon DIO, again primarily in male mice, which was similar in both WT and Duox1^−/− mice (Fig. 6B). To further address a potential impact of DUOX1 on glycemic control, we performed an intraperitoneal glucose tolerance test (IPGTT) after overnight fasting to normalize basal glucose levels (Fig. 6C). As expected, DIO resulted in glucose intolerance, indicated by greater peak glucose levels and delayed clearance, and increased IPGTT area-under-the-curve (AUC) (Fig. 6, D and E). Somewhat surprisingly, DIO-induced glucose intolerance was slightly but significantly attenuated in Duox1^−/− mice, in both males and females (Fig. 6, F and G). We also observed slightly increased glucose clearance during IPGTT in male nonobese (LFD-fed) Duox1^−/− mice compared with corresponding WT mice. To further address the potential impact of DUOX1 on glycemic control, we examined mRNA levels of various glucose transporters in either VAT or lung tissues. Consistent with earlier findings (45), the main insulin-sensitive glucose transporter in visceral adipose tissues (VAT), GLUT4, was downregulated during DIO, but this was similar in both WT and Duox1^−/− mice (Supplemental Fig. S6A). VAT mRNA levels of other facilitative glucose transporters, GLUT1 and GLUT2, were not significantly affected by either DIO or DUOX1, except that VAT GLUT2 mRNA tended to be enhanced in Duox1^−/− mice, which is interesting given the suggested importance of GLUT2 in glucose sensing and uptake (46). Similar analysis of lung tissues did not reveal any differences in GLUT mRNA expression due to either DIO or DUOX1 deficiency, except for a modest decrease in GLUT2 mRNA in response to DIO (Supplemental Fig. S6B). Collectively, these findings indicate that the apparent impact of DUOX1 on glycemic dysfunction could not be attributed to significant alterations in GLUT gene expression in either VAT or lung tissues. Intriguing recent studies demonstrated that glucose intolerance and insulin resistance in the context of DIO can be attributed to increased expression of EGFR and its ligand amphiregulin (AREG) in adipose tissue macrophages (47). Since we previously demonstrated a role for DUOX1 in macrophage Areg mRNA expression during appropriate stimulation (35), this prompted us to examine the potential contribution of DUOX1 on VAT Areg expression in the context of DIO. Indeed, HFD feeding led to the marked induction of Areg mRNA in VAT tissues, consistent with previous findings (47), and this was slightly but significantly attenuated in DUOX1-deficient mice (Fig. 6H).

Figure 6.

DUOX1 contributes to glucose intolerance by diet-induced obesity (DIO). A and B: analysis of nonfasted blood from low-fat diet (LFD)- or high-fat diet (HFD)-fed mice for glucose (A) or insulin (B). ***P < 0.001; ****P < 0.0001, n = 4–8. C: analyses of blood glucose of LFD and HFD groups after overnight fasting. D–G: intraperitoneal glucose tolerance test (IPGTT) was performed on overnight fasted LFD or HFD mice, and presented as combined analysis of IPGGT (D; n = 12) and area under the curve (AUC) (E), as well as separate analysis of male (F; n = 5 or 6) and female (G; n = 5 or 6) mice. *P < 0.05 compared with corresponding wild-type (WT) groups. H: effect of DIO on visceral adipose tissues (VAT) Areg mRNA expression. ***P < 0.001; ****P < 0.0001; 2-way ANOVA; n = 4 or 5.

Recent studies have indicated that short-term HFD feeding can induce rapid glucose intolerance in the absence of significant adiposity, which is distinct from metabolic alterations and glucose intolerance related to obesity due to prolonged HFD feeding (48). To determine whether glucose intolerance in the context of short-term HFD feeding is sufficient to induce alterations in acute HDM responses, we subjected mice to 3-wk HFD feeding before HDM challenge. Short-term (3-wk) HFD feeding induced only modest 20%–25% weight gain (Fig. 7A), compared with >100% weight gain in the case of prolonged (14-wk) feeding with HFD (Fig. 2C), and did not cause significant increases in BAL leptin (Fig. 7B), suggesting modest adiposity. However, nonfasting serum insulin was enhanced by approximately twofold (Fig. 7C), and glucose intolerance (measured using IPGTT) observed after 3-wk HFD feeding was comparable with that after long-term (14-wk) HDF feeding, based on peak glucose levels (Fig. 7D and Fig. 6D). In contrast to findings with long-term HFD feeding (Fig. 2), analysis of acute airway responses to HDM challenge after short-term (3-wk) HFD feeding did not reveal any differences in BAL IL-33 secretion (Fig. 7E) or lung tissue mRNA levels of type 2 cytokines (IL-5 and IL-13) or DUOX1 (Fig. 7, F–H). Hence, observed increases in airway DUOX1 and HDM-induced type 2 responses in the context of DIO are most likely due to metabolic alterations due to adiposity rather than by glucose intolerance per se.

Figure 7.

Short-term high-fat diet (HFD) feeding does not alter innate airway house dust mite (HDM) responses. C57BL/6NJ mice were subjected to 3-wk low-fat diet (LFD) or HFD feeding and analyzed for overall weight gain (A), bronchoalveolar lavage (BAL) leptin (B), serum insulin (C) or intraperitoneal glucose tolerance test (IPGTT) after overnight fasting (D). Mice subjected to 3-wk LFD- or HFD-feeding were challenged with HDM (1 h) and BAL fluid IL-33 (E) and lung tissue mRNA levels of Il5 (F), Il13 (G), and Duox1 (H) were analyzed. ***P < 0.001; Student’s t test; n = 2–4.

Association between DUOX1 and Leptin Production or Signaling

One potential mechanism by which DIO might affect allergen-induced type 2 inflammation is through the altered production of adipokines. For example, leptin (which is increased during DIO) can enhance Th2 and ILC2 responses (8), whereas adiponectin (which is reduced upon DIO), can suppress IL-33 signaling in ILC2 (9). As expected, DIO significantly enhanced leptin concentrations in serum (not shown) and also in BAL (Fig. 8A), and neither was affected by DUOX1 deficiency. To address the potential consequences of increased leptin, we examined leptin-mediated activation of airway basal cells in vitro (49, 50). Indeed, leptin induced dose-dependent increases in Il13 and Muc5ac mRNA in cultured airway basal cells from WT mice, which was significantly attenuated in airway basal cells from Duox1^−/− mice (Fig. 8B). Moreover, leptin (10 ng/mL) also significantly enhanced Duox1 mRNA expression in airway basal cells (Fig. 8C). Collectively, these findings indicate that the effects of DIO on type 2 inflammation may be due in part to increased leptin production, which can enhance epithelial DUOX1 and DUOX1-mediated activation of epithelial cytokine/mucin production.

Figure 8.

Involvement of DUOX1 in airway epithelial responses to leptin. A: increased bronchoalveolar lavage (BAL) leptin after 14-wk low-fat diet (LFD)- or high-fat diet (HFD) feeding. **P < 0.01; ***P < 0.001, n = 6. B and C: effect of airway basal cell stimulation with leptin (24 h) on mRNA levels of Il13 and Muc5ac (B) or Duox1 (C). *P < 0.05; **P < 0.01; 1- or 2-way ANOVA; n = 4–5 replicates from two independent experiments.

DISCUSSION

Adult asthma in association with obesity can be distinguished in two distinct endotypes: early-onset atopic asthma whose disease management or treatment is complicated by obesity in adulthood, and late-onset asthma in adulthood that occurs as a consequence of obesity (2, 3). The present studies aimed to address the former category, and examined the hypothesis that altered cytokine responses in obesity-associated allergic asthma (4, 5) may be related to altered activation of the NADPH oxidase DUOX1, which was previously found to be critical for innate epithelial responses to HDM allergen and type 2 inflammation during HDM-induced allergic airways disease (30, 31). Indeed, although previous studies have attributed enhanced HDM-induced type 2 or type 3 responses in the context of obesity to altered lung T helper cell or ILC profiles (6, 7), the present studies establish that (diet-induced) obesity is associated with increased epithelial expression of DUOX1, thereby resulting in amplified innate redox-dependent epithelial responses to airborne allergens such as HDM. Indeed, DUOX1 deletion did not only dramatically inhibit these various innate cytokine responses to HDM, but also largely eliminated DIO-induced increases in these responses. Thus, even though other mechanisms undoubtedly contribute to the overall effects of obesity on HDM-induced airway responses and allergic airway inflammation and remodeling, our present findings demonstrate that increased airway epithelial DUOX1 expression and activation contribute importantly to these worsened outcomes.

One important potential caveat in interpreting our findings with DUOX1-deficient animals is the possibility that DUOX1 deficiency also impacts HFD-induced weight gain and metabolic dysfunction. Indeed, even though DUOX1 deficiency did not significantly affect HFD-induced obesity and adiposity, it was found to contribute slightly to glycemic dysfunction as a result of HFD feeding. Such glycemic dysfunction and associated insulin resistance could conceivably also contribute to worsened asthma outcomes (11, 51). Indeed, insulin signaling in airway epithelial cells through the insulin receptor (INSR) is important for epithelial homeostasis and barrier function, and systemic hyperinsulinemia has been reported to promote airway remodeling and hyper-responsiveness (13). Moreover, a recent comparative study of epithelial gene expression patterns in human adult asthma indicated downregulation of various insulin target genes, which appeared to be most relevant for Th2-high molecular asthma endotypes (52), indicative of a contributing role for insulin resistance in asthma severity.

Altered expression or activation of NOX enzymes have been implicated in metabolic dysfunction associated with obesity or insulin resistance, although such relationships are complex and may involve proinflammatory effects due to increased expression of some NOX isoforms and enhanced ROS production, but also loss of some beneficial NOX signaling pathways involved in e.g., insulin signaling (32). For example, several studies have highlighted a contribution of NOX2-derived ROS in diet-induced adiposity and weight gain, which was linked to NOX2 activation in proinflammatory macrophages (53). In apparent contrast, genetic ablation of NOX4 was found to enhance diet-induced obesity (DIO) and insulin resistance, likely related to its beneficial roles in adipogenesis and insulin signaling (54). However, obesity and insulin resistance also result in elevated levels of adipose NOX4, and conditional deletion of NOX4 in adipocytes was actually found to delay insulin resistance induced by high-fat, high-sucrose diets, due to NOX4-dependent adipose tissue inflammation (55). The apparent contribution of DUOX1 to altered glycemic control, as observed in the present study, further adds to the complex nature of associations of NOX enzymes with metabolic (dys)function. Although DUOX1 expression is confined largely to epithelial cells at mucosal surfaces, in the lung, skin, etc., it also appears to be functional in some immune cell types (35, 56, 57), hence it will be difficult to determine how DUOX1 might affect systemic glycemic control. To begin to address this, we examined the potential association of DUOX1 with mRNA expression of several major glucose transporters in either the lung or in VAT (58), but observed no significant differences in the context of DUOX1 deficiency. However, NOX/DUOX-derived ROS could conceivably affect GLUT transporters critical for glucose uptake by oxidation of critical cysteines (59). Likewise, glucose homeostasis might also be affected by cysteine oxidation within SIRT6, a putative glucose homeostasis regulator (60). Another potential link between DUOX1 and glycemic control was suggested by an intriguing recent study, which highlighted the importance of enhanced EGFR signaling within adipose tissue macrophages in DIO-associated glycemic dysfunction and insulin resistance, due to enhanced expression of the EGFR ligand AREG (47). Indeed, we recently demonstrated that DUOX1 contributes to Areg expression in classically (M1) activated macrophages and in recruited macrophages in lung tissues during allergic airway inflammation, even though it didn’t contribute significantly to other features of macrophage M1 polarization (35). Accordingly, examination of VAT tissues did not show a significant effect of DUOX1 on adipose tissue macrophage M1 polarization in the context of DIO (Supplemental Fig. S1), but the marked increase in VAT Areg mRNA during DIO was significantly attenuated in DUOX1-deficient mice, consistent with our earlier observations (35), which may thus contribute to the slightly reduced glycemic dysfunction observed in these mice.

One intriguing observation in our studies is that hyperglycemia and hyperinsulinemia in the context of DIO was observed primarily in males, which might also be relevant for apparent sex-specific differences in type 2 inflammation (42). However, we did not observe a significant effect of sex on e.g., DUOX1 induction by DIO or HDM-induced IL-33 production, probably because animal numbers were not quite sufficient to reveal such differences. Thus, the apparent increase airway DUOX1 and allergen-induced type 2 inflammation in the context of DIO may not be strictly attributed to hyperglycemia or high insulin levels, which was more prominent in male animals. Also, attempts to address the effect of glucose intolerance in the absence of significant adiposity (using short-term HFD feeding) demonstrated no significant changes in acute innate airway responses to HDM (Fig. 7).

Based on the arguments earlier, we surmise that the impact of DIO on DUOX1 and airway responses to HDM must be due to alternative metabolic changes associated with adiposity, and therefore postulated that such increases in airway DUOX1 during adiposity may be related to altered production of adipokines such as leptin (61). Indeed, DIO induced significant increases in leptin in serum as well as BAL, which was similar in both WT and DUOX1-deficient mice, suggesting enhanced leptin signaling within the lung epithelium. Increased serum leptin as a result of obesity can worsen features of asthma (62), and airway epithelial cells express the leptin receptor (Ob-R) and can respond to leptin signaling with increased proinflammatory cytokine production (10, 63) and can also themselves produce leptin (64). We observed that in vitro stimulation of tracheal epithelial cells with leptin enhances DUOX1 expression and also enhances expression of IL-13 and MUC5AC, responses that were impaired in DUOX1-deficient cells, indicating a role of DUOX1 in such leptin signaling. Thus, enhanced airway DUOX1 expression in the context of obesity may, at least in part, be mediated by increased leptin production and leptin signaling in the airway epithelium. Such airway leptin signaling may also contribute to enhanced type 2 inflammation in the context of allergic airway inflammation, even though epithelial leptin receptor expression might actually be reduced during asthma (64). Thus, while leptin may enhance features of allergic asthma by various mechanisms (e.g., see Refs. 62 and 65), its ability to induce epithelial DUOX1 and augment DUOX1-mediated inflammatory responses represents one such mechanism. Contrasting the proinflammatory actions of leptin, adiponectin may have anti-inflammatory effects on the airway epithelial and its reduced levels in the context of obesity may therefore contribute to increased proinflammatory status in this case (66). However, our previous studies did not reveal significant decreases in BAL adiponectin in similar studies of DIO in mice (67).

In conclusion, our studies indicate that increased severity of allergic asthma in the context of obesity may at least in part be associated with airway DUOX1 function, and enhanced activation of type 2/type 17 inflammation and remodeling in response to airborne allergens such as HDM. It is worth noting that our chosen animal model of DIO before induction of HDM-induced allergic airway disease may not fully reflect the impact of obesity on early-onset allergic asthma, which typically precedes development of obesity and metabolic syndrome later in life. Nevertheless, our observations of increased DUOX1-mediated type 2 responses are consistent with various clinical findings indicating enhanced type 2 cytokine responses in obesity-associated asthma, which is associated with steroid-resistant disease (4, 5). In this regard, our findings may offer further justification of implicating DUOX1 as an appropriate alternative therapeutic target, particularly in cases of obesity-associated asthma in which other treatment strategies have proven ineffective.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Table S1 and Supplemental Figs. S1–S6: https://doi.org/10.6084/m9.figshare.21741071.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants R01 HL085646 and R01 HL138708 (to A.v.d.V.), R35 HL135828 (to Y.M.J.-H.), R01 HL142081 and R01 HL133920 (to M.E.P.), and NIH R01 HL136917 (to A.E.D. and V.A.).

DISCLOSURES

A.v.d.V. is listed on U.S. Patent No. 10143718, “Covalent Inhibitors of Dual Oxidase 1 (DUOX1),” issued on December 4, 2018.

AUTHOR CONTRIBUTIONS

A.v.d.V. conceived and designed research; A.H., M.H., and C.R.M. performed experiments; A.H., M.H., C.R.M., M.-C.J.L., L.C.C., J.L.A., M.G., and A.E.D. analyzed data; M.J.-C.L., M.G., V.A., Y.M.W.J.-H., M.E.P., A.E.D., and A.v.d.V. interpreted results of experiments; A.H. prepared figures; A.H. drafted manuscript; C.R.M., M.-C.J.L., L.C.C., J.L.A., V.A., Y.M.W.J.-H., M.E.P., A.E.D., and A.v.d.V. edited and revised manuscript; A.v.d.V. approved final version of manuscript.