Dual oxidase 2 is essential for house dust mite‐induced pro‐inflammatory cytokine production in human keratinocytes

House dust mites (HDMs) are known to trigger chronic inflammation through Toll‐like receptors (TLRs) and their signalling cascades. In this study, we found that TLR2 ligation by HDMs induced the activation of dual oxidase 2 (Duox2) and nuclear factor‐κB (NF‐κB), leading to the production of pro‐inflammatory cytokines in human keratinocytes. Stimulation of human keratinocytes with HDMs resulted in increases in interleukin‐8 (IL‐8) and chemokine (C–C motif) ligand 20 (CCL20) levels. However, pro‐inflammatory cytokine production was abolished in keratinocytes transfected with TLR2 siRNA, indicating that HDM‐induced cytokine production was mediated via TLR2 signalling. We also examined the function of Duox1/2 isozymes, which are primarily expressed in keratinocytes, in HDM‐mediated pro‐inflammatory cytokine production. Human keratinocytes transfected with control siRNA or Duox1 siRNA showed no inhibition of IL‐8 or CCL20 production in response to HDMs, whereas the silencing of Duox2 expression resulted in a failure to induce cytokine production. Moreover, the phosphorylation and nuclear localization of RelA/p65, a component of NF‐κB, were induced by HDMs in human keratinocytes. Transfection of human keratinocytes with TLR2 siRNA or Duox2 siRNA resulted in the complete abolishment of RelA/p65 nuclear localization in response to HDMs. Taken together, these results indicate that the HDM‐dependent TLR2‐Duox2 signalling axis indeed promotes NF‐κB activation, which induces IL‐8 and CCL20 production and mediates epidermal keratinocyte inflammation.

house dust mite; nuclear factor‐κB; reactive oxygen species; Toll‐like receptors; dual oxidase 2

Abbreviations

AD atopic dermatitis

CCL20 chemokine (C–C motif) ligand 20

Duox2 dual oxidase 2

HDM house dust mite

IL‐8 interleukin‐8

LPS lipopolysaccharide

NF‐κB nuclear factor‐κB

Nox NADPH oxidase

PBS phosphate‐buffered saline

ROS reactive oxygen species

RT room temperature

TLR Toll‐like receptors

Toll‐like receptors (TLRs) recognize the molecular patterns of microbes and activate pro‐inflammatory cell signalling cascades in the innate immune response and host defense [1] , [2] . Dermatophagoides farina, Dermatophagoides pteronyssinus and Euroglyphus maynei are common types of house dust mites (HDMs) in human habitats, and they trigger inflammatory events in the skin, such as atopic dermatitis (AD) [3] , [4] . HDMs reportedly induce the secretion of pro‐inflammatory cytokines, including interleukin (IL)‐1α, IL‐6, IL‐8, TNFα, GM‐CSF and VEGF, in the keratinocytes and fibroblasts of the skin [5] . Pro‐inflammatory cytokine production has been found to contribute to skin barrier disruption, leading to AD development [6] . Innate immunity is known to have an essential role in the HMD‐induced inflammatory process [3] , [4] . HDM‐induced innate immunity is activated by the TLR, C‐type lectin receptor, NOD‐like receptor and proteinase‐activated receptor signalling pathways [3] . A recent report has indicated that HDMs activate the nucleotide‐binding oligomerization domain, leucine‐rich repeat and pyrin‐domain containing 3, which in turn stimulate IL‐1β and IL‐18 secretion, promoting AD development [7] .

Recently, many reports have indicated that NADPH oxidase (Nox) isozymes play essential roles in pro‐inflammatory cell signalling and host defense [8] , [9] , [10] . Reactive oxygen species (ROS) generation as a host defense mechanism has been extensively studied for gp91phox/Nox2 in phagocytic cells, resulting in the identification of six homologues of gp91phox (Nox1, Nox3, Nox4, Nox5, Duox1 and Duox2) in a variety of cell types [8] , [9] . In particular, Duox1/2 activation contributes to the regulation of various physiological functions, such as host defense and innate immunity [11] , [12] . The epithelial Duox/ROS system may play an important role in host defense because nasal and gut mucosal epithelia, which primarily express Duox1/2, come into contact with various microbes [11] , [12] . Yoo et al. reported that β‐glucan/TLR2‐mediated Duox2/ROS regulate innate immunity in the nasal mucosa and trigger HDM‐induced nasal inflammation [13] . In addition, HDMs reportedly stimulate pro‐inflammatory cytokine production in inflamed keratinocytes [5] , [7] . However, the molecular connection between HDM‐mediated pro‐inflammatory cytokine production and ROS generation in epidermal inflammation remains unclear. In this study, we demonstrated that HDMs stimulated ROS generation by activating the TLR2‐Duox2 cascade. Activation of this cascade then mediated nuclear factor‐κB (NF‐κB) activation, leading to the production of pro‐inflammatory cytokines, such as IL‐8 and CCL20, in human keratinocytes.

Neonatal human epidermal keratinocytes derived from neonatal foreskin were purchased from Lonza (Muenchensteinerstrasse, Basel, Switzerland) and cultured in KBM medium with KGM2 growth supplements, containing insulin, human epidermal growth factor, bovine pituitary extract, hydrocortisone, epinephrine, transferrin and gentamicin/amphotericin B, which were also purchased from Lonza. Cells were serially diluted at 70–80% confluence, and experiments were conducted using subconfluent cells at passage two or three during the proliferative growth phase.

Neonatal human epidermal keratinocytes were transfected with 50 nm of ON‐TARGETplus SMARTPOOL siRNA (Non‐targeting #2: #D‐001810‐02‐20; human TLR2: #L‐005120‐01; human DUOX1: #L‐008126‐00; and human DUOX2: #L‐008324‐00) according to the manufacturer's protocol (Thermo Fisher Scientific, Waltham, MA, USA). In brief, cells were seeded in 60‐mm dishes and transfected with siRNA using RNAiMAX (Invitrogen, Carlsbad, CA, USA) and OPTI‐MEM (Invitrogen) for 16 h. The medium was then changed to KBM medium containing all supplements.

Total RNA was isolated using TRIzol reagent (Invitrogen) according to the manufacturer's instructions. Two micrograms of RNA was reverse‐transcribed into cDNA using SuperScript III reverse transcriptase (Invitrogen). Quantitative real‐time TaqMan RT‐PCR (qRT‐PCR; Applied Biosystems, Foster City, CA, USA) was performed to determine the expression levels of the selected target genes. The cycling conditions included a denaturing step at 95°C for 10 min and 50 cycles at 95°C for 15 s, followed by annealing and elongation at 60°C for 1 min. The TaqMan probes (Applied Biosystems) used for real‐time qRT‐PCR were NOX1 (Hs00246589_m1), NOX2 (Hs00166163_m1), NOX3 (Hs00210462_m1), NOX4 (Hs00276431_m1), NOX5 (Hs00225846_m1), DUOX1 (Hs00213694_m1), DUOX2 (Hs00204187_m1), TLR2 (Hs01872448_s1) and TLR4 (Hs00152939_m1). Human GAPDH (4333764F) was also amplified to normalize variations in cDNA quantities from different samples.

House dust mites (D. pteronyssinus) were obtained from the Arthropods of Medical Importance Bank (Yonsei University, Seoul, Korea). Mites were cultured in an insect‐rearing facility at the AMIB, as described previously [13] , [14] , [15] , [16] . Briefly, 30 g of frozen dust mites was pulverized in liquid nitrogen, and defatted samples in 200 ml of a 1:1 volume of ethyl ether/ethyl acetate were extracted with slow stirring at 4°C overnight in a 100 mm phosphate‐buffered saline (PBS) solution (pH 7.4). The extracts were then centrifuged at 10 000 g for 30 min at 4°C, and the supernatants were filtered through a 0.45‐μm filter. The final concentration of the HDM extracts was 1 mg/ml. Endotoxin levels were measured using the limulus amebocyte lysate QCL‐1000 test (Lonza, Basel, Switzerland) (Lonza), and the average endotoxin level was determined to be 4.661 EU/mg (approximately 0.5 ng/ml). (1→3)‐β‐glucans were also measured using the alternative limulus amebocyte lysate assay system, and their concentration in the HDM extracts was found to be approximately 4.5 ng/ml in our study.

After stimulation of the confluent cells, they were washed with Hanks’ balanced salt solution and incubated for 10 min in the dark at 37°C in the same solution containing 10 μm 2′,7′‐dichlorofluorescein diacetate (DCF‐DA; Molecular Probes, Eugene, OR, USA). Cells were then examined using a laser‐scanning confocal microscope (model LSM 510; Carl Zeiss, Oberkochen, Germany) equipped with an argon laser tuned to an excitation wavelength of 488 nm, an LP505 emission filter (515–540 nm) and a Zeiss Axiovert 100× objective lens. Images were digitized and stored at a resolution of 512 by 512 pixels. Five groups of cells were randomly selected from each sample. Then, the mean relative fluorescence intensity for each group of cells was measured using a Zeiss vision system (LSM 510, version 2.3), and this value was averaged for all groups. All experiments were repeated at least three times.

We performed ELISAs for IL‐8 (D8000C) and CCL20 (DM3A00) with kits obtained from R&D Systems (Minneapolis, MN, USA), according to the manufacturer's instructions. Cells were seeded in a 60‐mm dish at 106 cells per well and treated with HDM extracts. After 24 h of incubation, cell‐free supernatants were harvested and measured by ELISA.

Cells were chilled in lysis buffer [50 mm Tris‐HCl (pH 7.4), 1% Triton X‐100, 0.5% NP‐40, 150 mm NaCl, 1 mm EDTA, 0.1 m 4‐(2‐aminoethyl) benzenesulfonyl fluoride, 1 mm Na3VO4, 1 mm sodium fluoride, 1 μg/ml aprotinin, 1 μg/ml leupeptin and 10% glycerol] for 30 min and then sonicated (at 10% amplitude for 1 s in a Branson sonifier equipped with a microtip), followed by centrifugation for 30 min at 14 000 rpm in a microcentrifuge. Next, the samples were boiled in 5× sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS‐PAGE) sample buffer and subjected to 10% SDS‐PAGE. The resolved proteins were electrotransferred to a nitrocellulose membrane. Then, the membrane was immunoblotted with an NF‐κB p65 antibody (#3034; Cell Signaling Technology, Danvers, MA, USA), followed by a horseradish peroxidase‐conjugated goat anti‐rabbit IgG antibody. Bands were visualized by chemiluminescence (Fujifilm LAS‐3000; Fujifilm Inc., Tokyo, Japan).

Neonatal human epidermal keratinocytes were seeded onto 18‐mm glass coverslips, cultured for 24 h and then incubated for 16 h in serum‐free medium. Cells were stimulated with HDMs (50 μg/ml) for the indicated times, washed with PBS, fixed with 3.5% paraformaldehyde in PBS for 10 min at room temperature (RT) and permeabilized in 0.5% Triton X‐100. Non‐specific sites were blocked by treating the cells with PBS containing 0.5% bovine serum albumin for 1 h at RT. The cells were incubated with a primary anti‐NF‐κB p65 antibody (#3034; Cell Signaling Technology) in PBS overnight at 4°C, washed with PBS and then incubated with a secondary antibody (Alexa Fluor® 594 goat anti‐rabbit IgG, A‐11037; Molecular Probes) for 1 h. Next, the samples were washed with PBS, stained with DAPI for 10 min and then mounted with mounting solution (Sigma‐Aldrich Co., St. Louis, MO, USA) after washing with PBS. Images were recorded using a confocal laser‐scanning microscope (Carl Zeiss Meta 510; Carl Zeiss).

Human Cytokine Array Panel A (ARY005; R&D Systems) and a Human Chemokine Array Kit (ARY017; R&D Systems) were used to assay over 36 cytokines and 31 chemokines, respectively, in supernatants of sorted cell cultures. The array membranes were incubated in blocking buffer for 1 h at RT. Then, 1.5 ml of the sample/antibody mixture was added per well, followed by incubation overnight at 2–8°C on a rocking platform shaker. The membranes were washed three times in wash buffer at RT. Next, streptavidin‐HRP in array buffer was added, and the membranes were incubated for 30 min at RT. The membranes were washed again, followed by addition of Chemi Reagent Mix for 1 min. Then, the membranes were visualized using an LAS 3000 chemiluminescence imaging system (Fujifilm Inc.).

All of the data are presented as the mean ± SE. Significant differences between treatment groups were identified using the t‐test. P values of <0.05 were considered statistically significant.

To examine the pro‐inflammatory cytokine production in human keratinocytes, supernatants from human keratinocytes cultured with or without HDM stimulation were assessed using chemokine and cytokine arrays (ARY017; R&D Systems). We found that only IL‐8 expression was increased in response to HDMs in these cells (Fig. [NaN] a). To validate this array result, we measured the quantitative mRNA and protein levels of IL‐8. Stimulation of keratinocytes by HDMs resulted in significant increase in IL‐8 expression (Fig. [NaN] b) and secretion (Fig. [NaN] c). A previous report has suggested that HDM stimulation increases CCL20 chemokine expression [13] , [17] . Therefore, we measured CCL20 production in human keratinocytes in response to HDMs. Although we failed to detect increased CCL20 levels in culture supernatants from human keratinocytes activated by HDMs in the chemokine and cytokine arrays, we found that its quantitative mRNA and secretion levels in these cells were increased in response to HDMs (Fig. [NaN] d,e). The CCL20 protein expression level was likely significantly lower than that of IL‐8; thus, it could not be detected in the chemokine or cytokine array. These results indicate that HDMs induce the production of pro‐inflammatory cytokines, such as IL‐8 and CCL20, in human keratinocytes.

House dust mites HDMs contain β‐glucan, a TLR2 agonist, and lipopolysaccharide (LPS), a TLR4 agonist [13] , [18] , [19] . In this experiment, we used β‐glucan‐deleted HDMs (HDMΔβ‐glucan), LPS‐defective HDMs (HDMΔLPS) and protease‐inactivated HDMs (HDMΔprotease) to verify the role of the TLR signalling cascade in pro‐inflammatory cytokine production. HDM β‐glucan (HDMΔβ‐glucan) degradation and LPS (HDMΔβ‐glucan) inactivation were achieved by pretreatments with β‐glucanase and polymyxin B, respectively. Moreover, HDM protease (HDMΔprotease) was inactivated by a heat pretreatment. Incubation of human keratinocytes with either HDMΔprotease or HDMΔLPS resulted in increases in IL‐8 and CCL20 production, whereas HDMΔβ‐glucan failed to stimulate the expression of these two pro‐inflammatory cytokines (Fig. [NaN] f,g). Furthermore, TLR2 expression was downregulated following transfection of human keratinocytes with TLR2‐specific siRNA (Fig. [NaN] j), verifying the role of the TLR2‐dependent signalling pathway in pro‐inflammatory cytokine expression. Control siRNA transfection displayed no inhibitory effect on IL‐8 or CCL20 production in response to HDMs; however, the silencing of TLR2 expression resulted in a failure to induce cytokine production in human keratinocytes (Fig. [NaN] h,i). These results indicate that HDM‐induced IL‐8 and CCL20 production is mediated by TLR2‐dependent cell signalling.

A previous report has suggested that HDM‐induced pro‐inflammatory cytokine production is required for ROS generation in the upper and lower airways [13] . To examine this possibility in human keratinocytes, we measured the effect of DPI, an Nox inhibitor, on ROS levels. Pretreatment of human keratinocytes with DPI resulted in abolished ROS generation in response to HDMs (Fig. [NaN] a). Next, we examined the expression levels of Nox isozymes in human keratinocytes to better understand the response of the ROS‐generating system of the HDM‐stimulated epidermis. The real‐time qRT‐PCR results indicated that Duox1 and Duox2 were the predominant Nox isozymes in human keratinocytes (Figure S1). Other Nox isozymes were barely detectable (Figure S1). These results suggest that Duox isozymes might be responsible for HDM‐induced ROS generation in human keratinocytes. To explore the roles of Duox1 and Duox2 in HDM‐mediated ROS generation, we used human keratinocytes transfected with Duox1 siRNA or Duox2 siRNA. The transfected cells exhibited a marked reduction (70%) in the level of each targeted mRNA (Fig. [NaN] b). The culture supernatants from keratinocytes transfected with control siRNA or Duox1 siRNA displayed increased IL‐8 production in response to HDMs (ARY005, 36 chemokines and cytokines). In contrast, human keratinocytes with silenced Duox2 expression did not exhibit induction of IL‐8 production (Fig. [NaN] c). These results indicate that HDMs stimulate the expression of pro‐inflammatory cytokine, IL‐8 in human keratinocytes through Duox2 activation.

We confirmed that HDM‐induced upregulation of IL‐8 and CCL‐20 was significantly reduced in human keratinocytes transfected with Duox2‐specific siRNA (Fig. [NaN] d). In addition, we also found that knockdown of TLR2 or Duox2 resulted in significantly decreased ROS generation in response to HDMs (Fig. [NaN] e). This result clearly indicates that ROS generation via HDM‐mediated Duox2 activation occurs through TLR2.

Next, we examined whether HDMs stimulate NF‐κB to promote IL‐8 and CCL20 production through the TLR2‐Duox2 axis in human keratinocytes. HDM‐mediated stimulation of human keratinocytes resulted in increased serine phosphorylation of RelA/p65, which is an NF‐κB component (Fig. [NaN] a). The nuclear localization of RelA/p65 appears to be a hallmark event of agonist‐dependent NF‐κB activation. We found that RelA/p65 nuclear localization was induced by HDM stimulation (Fig. [NaN] b,c). Moreover, pretreatment with an NF‐κB inhibitor (481407, Calbiochem) significantly inhibited IL‐8 and CCL20 production in response to HDMs in human keratinocytes (Fig. [NaN] d,e). These results suggest that HDMs induce pro‐inflammatory cytokine production through NF‐κB activation. However, the NF‐κB inhibitor did not affect ROS generation in response to HDMs in these cells, indicating that Nox activation occurs upstream of NF‐κB activation in the TLR2 signalling cascade (Figure S2).

To validate the effect of the HDM‐TLR2‐Duox2 cascade on NF‐κB activation, we transfected human keratinocytes with either TLR2‐ or Duox2‐specific siRNA and then analysed p65 nuclear localization in response to HDMs. TLR2‐ or Duox2‐specific siRNA‐transfected human keratinocytes exhibited dramatically decreased p65 nuclear localization in response to HDMs compared with control siRNA‐transfected human keratinocytes (Fig. [NaN] a,b). These results clearly demonstrate that HDMs stimulate the TLR2‐Duox2 cascade, which in turn induces NF‐κB activation and IL‐8 and CCL20 production in inflamed human keratinocytes.

Reactive oxygen species ROS have recently been recognized as second messengers in receptor‐mediated cell signalling [8] , [9] , [20] . The ligation of receptors with agonists, including growth factors, hormones, cytokines and microbes, induces ROS generation, leading to the mediation of various biological events, including cell growth, differentiation, and inflammation. Cellular inflammation and pro‐inflammatory cytokine production are known to be associated with ROS generation through Nox activation [21] . Because skin keratinocytes exclusively express the Duox1/2 isozymes, we analysed the role of Duox1/2 activation in the generation of ROS and the production of pro‐inflammatory cytokines, such as IL‐8 and CCL20, in response to HDM stimulation. Duox2 activation by HDM‐TLR2 was found to play an essential role in ROS generation and in IL‐8 and CCL20 production (Fig. [NaN] ). Previous reports have suggested that injury‐induced Nox2 activation stimulates the production of IL‐8, a neutrophil chemoattractant that may serve to recruit inflammatory cells in response to local and temporal changes in the redox level [22] . Our results provide mechanistic insights into redox changes, such as Duox2 activation and pro‐inflammatory cytokine production that occur in skin keratinocytes (Fig. [NaN] c).

Skin barrier dysfunction results in the occurrence of AD [23] , [24] . A previous report has suggested that stimulating keratinocytes from AD patients with Pam3Cys, which is a TLR2 agonist, result in increased IL‐8 and CCL20 production, which might contribute to skin barrier dysfunction [25] . Post et al. have reported that the HDM‐mediated activation of human bronchial epithelial cells results in intra‐cellular calcium mobilization and increased CCL20 cytokine production. These changes play important roles in barrier dysfunction and cellular inflammation [17] .

In the present study, we demonstrated that HDM‐dependent pro‐inflammatory cytokine production occurs through the TLR2‐Duox2‐NF‐κB cascade (Fig. [NaN] c). Pro‐inflammatory cytokines, including IL‐8 and CCL20, play an important role in chronic inflammation. Moreover, Ryu et al. [13] have provided experimental evidence that HDM‐dependent pro‐inflammatory cytokine production contributes to allergic rhinitis and asthma through activation of the TLR2‐Duox2 axis. Epidermal keratinocytes in AD patients highly express TLR2, and our data consistently indicated that HDM‐induced pro‐inflammatory cytokine production (IL‐8 and CCL20) is regulated by TLR2 signalling (Fig. [NaN] ). Moreover, a previous report has indicated that TLR2 induces the expression of CCL20, CCL2, MMP9 and IL‐8 in human keratinocytes [26] . Previous reports and our present data indicate that HDM‐induced pro‐inflammatory cytokine production may be related to disruption of the skin barrier structure and to skin barrier dysfunction.

Reactive oxygen species ROS have been reported to regulate the phosphorylation of cytosolic proteins by inhibiting protein tyrosine phosphatases [27] , [28] . NF‐κB, a well‐known transcription factor regulated by the redox level [29] , [30] , binds to the IkB protein, which can be phosphorylated by IkB kinase. An increase in redox level‐dependent IkB phosphorylation is recognized by E3 ligase, and the protein is subsequently subjected to proteasomal degradation, allowing for the release and nuclear localization of NF‐κB for redox‐sensitive protein expression [31] , [32] . Several studies reported that redox‐mediated NF‐κB activation is involved in pro‐inflammatory cytokine production [33] , [34] , [35] . Based on this notion, we conclude that β‐glucan in HDMs stimulates the TLR2‐Duox2 cascade, resulting in ROS generation (Fig. [NaN] c). Taken together, these data provide mechanistic insights, demonstrating that HDM‐mediated redox generation in epidermal inflammation plays an important role in pro‐inflammatory cytokine production.

This work was supported by the National Research Foundation of Korea (NRF) (grant No. 2012R1A5A1048236), by the Bio & Medical Technology Development Program (No. 2012M3A9B4028785) and by a Redoxomics Grant (grant 2012M3A9C5048708) funded by the Ministry of Science, ICT & Future Planning.

EB Ko and H Choi designed the research study. EB Ko, H Choi, KN Park and JY Park performed the research. EB Ko and H Choi analysed the data. TR Lee, DW Shin and YS Bae wrote the manuscript. All authors approved this manuscript.

The authors have declared no conflicting interests.

Graph: House dust mites ( HDM s) stimulate pro‐inflammatory cytokine production through Toll‐like receptors 2 ( TLR 2) signalling in human keratinocytes. (a) Pro‐inflammatory cytokine production profiles of human keratinocytes with or without HDM stimulation, as determined using a human chemokine array. *The circles denote reference spots. (b–e) HDM s triggered cytokine production in normal human keratinocytes in dose‐ and time‐dependent manners. Quantitative m RNA and protein levels of interleukin‐8 ( IL ‐8) (b, c) and chemokine (C–C motif) ligand 20 ( CCL 20) (d, e). The data are presented as the mean ±  SE of three independent experiments; * P  < 0.01 and ** P  < 0.05. (f, g) IL ‐8 and CCL 20 expression following treatment with HDM s, HDM Δprotease, HDM Δβ‐glucan or HDM Δ LPS (50  μ g/ml each). (h, i) IL ‐8 and CCL 20 secretion following HDM treatment in human keratinocytes transfected with control or TLR 2‐specific si RNA using RNA i MAX. All of the results are presented as the mean ±  SE and are representative of three independent experiments; * P  < 0.01 and ** P  < 0.05. (j) TLR 2 m RNA expression was assessed by real‐time q RT ‐ PCR as described in.

Graph: image_n/exd12808-fig-0001.png

Graph: Toll‐like receptors 2 ( TLR 2) and dual oxidase 2 (Duox2) are required for house dust mite ( HDM )‐induced reactive oxygen species ( ROS ) generation and pro‐inflammatory cytokine production in human keratinocytes. (a) Effects of a NADPH oxidase ( N ox) inhibitor on HDM ‐induced ROS generation. Cells were pretreated with the N ox inhibitor DPI (10  μm ) for 30 min, and ROS generation was monitored after 10 min of HDM (50  μ g/ml) treatment. The data are presented as the mean ±  SE ( n  = 3); * P  < 0.05. (b) D uox1 and D uox2 m RNA expression levels were assessed by real‐time q RT ‐ PCR , as described in. (c) HDM s trigger IL ‐8 expression via D uox2 in human keratinocytes. Primary human keratinocytes were transfected with either D uox1 or D uox2 or control si RNA using RNA i MAX. Following HDM (50  μ g/ml) treatment, pro‐inflammatory cytokine production profiles were determined in human keratinocytes using a human cytokine array. *The circles denote reference spots. (d) interleukin‐8 ( IL ‐8) and chemokine (C–C motif) ligand 20 ( CCL 20) secretion following HDM challenge in human keratinocytes transfected with control or D uox2‐specific si RNA using RNA i MAX. All of the results are presented as the mean ±  SEM and are representative of three independent experiments. ** P < 0.01. (e) HDM s stimulate H2O2 generation via TLR 2 and D uox2 in human keratinocytes. Human keratinocytes were transfected with control or Toll‐like receptors 2 ( TLR 2)‐ or D uox2‐specific si RNA using RNA i MAX. Following HDM (50  μ g/ml) treatment, H2O2 generation was monitored by confocal microscopic analysis of 2′,7′‐dichlorofluorescein ( DCF ) fluorescence. The data are presented as the mean ±  SE ( n  = 3). * P <0.05 TLR 2 and D uox2 m RNA expression levels were assessed by real‐time q RT ‐ PCR , as described in.

Graph: image_n/exd12808-fig-0002.png

Graph: Nuclear factor‐κB ( NF ‐κ B ) is required for the house dust mite ( HDM )‐induced innate immune response in human keratinocytes. (a) NHEK s were incubated with HDM s (50  μ g/ml) for the indicated times. Then, lysates were subjected to immunoblot analysis with NF ‐κ B p65 and phospho‐ NF ‐κ B p65 antibodies. (b) Cells were incubated with HDM s (50  μ g/ml) for the indicated times. Samples were stained with an NF ‐κ B p65 antibody (red) and DAPI (blue). Fluorescent images were visualized using a confocal microscope (600×). (c) The percentage of fluorescent cells in 10 random fields was determined. Data were analysed using Student's t ‐test. (d, e) interleukin‐8 ( IL ‐8) and chemokine (C–C motif) ligand 20 ( CCL 20) secretion following HDM treatment in human keratinocytes pretreated with an NF ‐κ B activation inhibitor (10  μm ). All of the results are presented as the mean ±  SE and are representative of three independent experiments; * P  < 0.01, ** P  < 0.05 and *** P  < 0.005.

Graph: image_n/exd12808-fig-0003.png

Graph: Nuclear factor‐κB ( NF ‐κ B ) activation is dependent on Toll‐like receptors 2 ( TLR 2) and dual oxidase 2 ( D uox2) expression in human keratinocytes. (a) Cells were transfected with TLR 2‐ and D uox2‐specific si RNA and stimulated with HDM s (50  μ g/ml) for 20 min. Samples were stained with an NF ‐κ B p65 (red) antibody and DAPI (blue). Fluorescent images were visualized using a confocal microscope (600×). (b) The percentage of fluorescent cells in 10 random fields was determined. Data were analysed using Student's t ‐test; * P  < 0.005. (c) Proposed model for HDM stimulation of the TLR 2‐ D uox2 cascade, resulting in reactive oxygen species ( ROS ) generation and pro‐inflammatory cytokine production.

Graph: image_n/exd12808-fig-0004.png

Graph: Figure S1. Expression of Nox isoforms in primary human keratinocytes. Figure S2. Effects of a NF‐κB activation inhibitor on HDM‐induced ROS generation.