Six-week-old female BALB/c mice were purchased from SLC Inc. (Hamamatsu, Japan). The mice were housed with 5 mice/cage in a laminar air flow room maintained at a temperature of 22±2°C, a relative humidity of 55±5% and a 12 h light:dark cycle throughout the study. The care and treatment of the mice were in accordance with the guidelines established by the Public Health Service Policy on the Humane Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee of Kyungpook National University.

D. kaki calyx used in this study was purchased from the Oriental drug store Bohwa Dang (Jeonju, Korea) and identified by Dr D.K. Kim at the College of Pharmacy, Woosuk University. A voucher specimen (no. WSP-15-098) was deposited at the Herbarium of the College of Pharmacy, Woosuk University. The sample was extracted with purified water at 70°C for 5 h (2 times) in a water bath. Then the extract was filtered, lyophilized, and then kept at 4°C. The yield of dried extract from starting crude materials was ~10.1%.

For the animal experiments, the dried residue was dissolved in phosphate-buffered saline (PBS). DFE (Greer Laboratories, Lenoir, NC, USA) and 2,4-dinitrochlorobenzene (DNCB) were used as antigen and hapten for the induction of AD-like skin lesions, respectively. All other reagents were purchased from Sigma (St. Louis, MO, USA) unless otherwise stated. DFE was dissolved in PBS containing 0.5% Tween-20. DNCB was dissolved in an acetone/olive oil (1:3) solution. Recombinant human tumor necrosis factor-α (TNF-α) and IFN-γ were purchased from R&D Systems (Minneapolis, MN, USA).

A human keratinocyte cell line, HaCaT, was maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics (100 U/m penicillin G, 100 µg/ml streptomycin) at 37°C under 5% CO₂. Passages 3–6 were used throughout the study.

AD-like lesions were induced by DFE and DNCB according to previous studies (4,5). The schematic experimental procedure is described in Fig. 1A. Female BALB/c mice were divided into 6 groups (n=5): vehicle, DFE/DNCB plus vehicle, DFE/DNCB plus AEDKC (1, 10 and 100 mg/kg), or dexamethasone (Dexa, 1 mg/kg). Mice were anesthetized with ketamine and the surfaces of both ear lobes were very gently stripped 4 times with surgical tape (Nichiban, Tokyo, Japan). Then, 20 µl of DNCB (1%) was painted on each ear and then with 20 µl of DFE (10 mg/ml) 4 days later. DFE/DNCB exposure was repeated once a week rotationally for 4 weeks. Two weeks after the first induction, tail bleeding was performed to evaluate the serum IgE level. After the confirmation of AD, as indicated by the IgE level, AEDKC was orally administered until the end of the 4-week induction (5 times/week). Ear thickness was measured the following day at the same time after DFE or DNCB application using a dial thickness gauge (Mitutoyo, Co., Tokyo, Japan).

On day 28, blood samples were collected by orbital puncture. After the blood had clotted at room temperature, it was centrifuged at 400 x g for 15 min at 4°C, and the serum was isolated. The serum was stored at −80°C for additional analysis. The ear of each mice was removed and subjected to histopathological analysis. Serum IgE and IgG2a levels were measured using an enzyme-linked immunosorbent assay (ELISA) kit (BD Biosciences, Oxford, UK) according to the manufacturer's instructions. For the detection of DFE-specific IgE, 96-well plates (Nunc, Wiesbaden, Germany) were coated with 10 mg DFE in PBS. The DFE-specific IgE level was indicated by the OD value.

The ears were fixed with 10% formaldehyde and embedded in paraffin. Sections (5 µm) were stained with hematoxylin and eosin (H&E) and toluidine blue (TB). To measure Infiltrated lymphocytes, thickening of the epidermis, and fibrosis in the dermis, skin sections were stained with H&E and the stained fields were observed by microscopy. To measure mast cell infiltration, skin sections were stained with TB, and the number of mast cells in the 5 sites chosen at random was counted. Eosinophils were counted in a blinded manner in 10 high-power fields at a magnification of ×400. Epidermal and dermal thickness of the H&E-stained sections was analyzed under a magnification of ×200. Thickness was measured in 5 randomly chosen fields from each sample.

Histamine content was measured following the o-phthaldialdehyde spectrofluorometric procedure of a previous study (25). The blood from mice was centrifuged at 400 x g for 15 min, and the serum was diluted with PBS and withdrawn to measure the histamine content. Fluorescence intensity was measured using 355-nm excitation and 450-nm filters and the fluorescence spectrometer LS-50B (Perkin-Elmer, Norwalk, CT, USA).

To detect the expression of cytokines, qPCR was performed using the Thermal Cycler Dice TP850 (Takarabio Inc., Shiga, Japan) according to the manufacturer's protocol. HaCaT cells (1×10⁵ cells/24-well plate) were pretreated with AEDKC for 1 h, and then stimulated with TNF-α (10 ng/ml) and IFN-γ (10 ng/ml) for 6 h. Total cellular RNA was isolated from cells as described in a previous study (4). Briefly, 1 µl of cDNA (100 ng), 1 µl of a sense and antisense primer solution (0.4 µM), 12.5 µl of SYBR Premix Ex Taq (Takarabio Inc), and 9.5 µl of nuclease-free water were mixed together to obtain a final 25 µl reaction mixture in each reaction tube. The conditions for PCR were similar to those of a previous study (4). The normalization and quantification of mRNA expression were performed using TP850 software supplied by the manufacturer.

Samples for western blot analysis were prepared in accordance with a previous study (26). Briefly, HaCaT cells (2×10⁶ cells/6-well plate) were pretreated with AEDKC for 1 h and then stimulated with TNF-α (10 ng/ml) and IFN-γ (10 ng/ml) for 30 min to activate signal transducer and activator of transcription 1 (STAT1) and nuclear factor-κB (NF-κB). Cells were rinsed once with ice-cold PBS, and total cell lysates were collected in 100 µl of lysis buffer. The lysates were spun in a micro-centrifuge for 20 min at 4°C, and the supernatant was collected. Proteins were subjected to 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to nitrocellulose membranes. The membranes were stained with reversible Ponceau S to ensure equal loading of samples onto the gels. Nuclear and cytosolic p65 NF-κB and IκBα were assayed using anti-NF-κB (p65; sc-109) and anti-IκBα (sc-371) antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), respectively. The phosphorylation of STAT1 was detected using anti-STAT1 (#9172) and anti-phospho-STAT1 (#9167) antibodies (Cell Signaling Technology, Beverly, MA, USA). Anti-β-actin (sc-8432; mouse monoclonal; 1:1,000) was from Santa Cruz Biotechnology, Inc. Immunodetection was conducted using Supersignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific, Waltham, MA, USA).

Statistical analyses were performed using Prism 5 (GraphPad Software, San Diego, CA, USA). Treatment effects were analyzed using one-way analysis of variance followed by Dunnett's test. A p-value <0.05 indicates a statistically significant difference.

To investigate the efficacy of AEDKC on AD-like skin lesions, a DFE/DNCB-induced AD-like model was used. During the induction period, the ear swelling of mice was measured after 24 h of DFE or DNCB induction (Fig. 1B). The tendency of ear swelling of mice in each group was similar until after 2 weeks. After 19 days of AD induction, oral administration of AEDKC considerably reduced ear thickness. DFE/DNCB treatment during the induction period (4 weeks) evoked severe AD-like skin lesions (Fig. 2A). Mouse ears became red and swollen after DFE/DNCB exposure. However, oral treatment of AEDKC relieved these symptoms compare to the AD mice. Dexamethasone (Dexa) was used as a positive control drug. Oral administration of AEDKC five consecutive days a week during 3 weeks did not alter the body weight of mice (data not shown), indicating that AEDKC exhibited no toxic effects.

Histological analysis showed that AEDKC treatment significantly suppressed erythema and infiltration of acute inflammatory cells compared with the AD mice (Fig. 2B–D). Epidermis thickening is regarded as an important factor that contributes to ear swelling (22). Compared to the AD mice, AEDKC considerably decreased DFE/DNCB-induced epidermal and dermal thickness (Fig. 2B) and infiltration of eosinophils (Fig. 2C). Mast cell-derived inflammatory mediators and histamine contribute to itching and inflammation in AD (7). Thus, mast cell infiltration into the AD site and serum histamine level were measured; AEDKC attenuated both mast cell infiltration (Fig. 2D) and serum histamine (Fig. 3A).

We previously reported a substantial increase in serum immunoglobulin and cytokines in DFE/DNCB-induced AD mice (4). To distinguish the role of AEDKC on the Th1 and Th2 response, we examined individually the serum levels of IgG2a and IgE (total and DFE-specific). Compared with the AD mice, the levels of total IgE, DFE-specific IgE, and IgG2a were markedly decreased in the serum of mice treated with AEDKC (Fig. 3B–D).

After establishing the inhibitory effect of AEDKC on AD mice, the keratinocyte model was used to ascertain the molecular mechanism and biological function of AEDKC. Keratinocytes have been broadly used to imitate the AD environment in vitro (27). They exhibit a similar immune response during the development of skin disorders, such as AD-like skin lesions (28). First, we evaluated the cytotoxicity of AEDKC by exposing HaCaT cells to various concentrations of AEDKC for 24 h. In the MTT assay, AEDKC did not exert cytotoxicity at concentrations up to 1,000 µg/ml in the keratinocytes (data not shown). To examine the effect of AEDKC on pro-inflammatory cytokines and a chemokine, HaCaT cells were pretreated with AEDKC for 1 h, followed by the stimulation with TNF-α/IFN-γ for 6 h. The results of qPCR indicated that AEDKC inhibited TNF-α/IFN-γ-induced gene expression of TNF-α, IL-1β, IL-6 and CCL17 in the HaCaT cells (Fig. 4).

Thereafter, the regulatory effect of AEDKC on the expression of pro-inflammatory cytokines and a chemokine was examined. To establish the mechanism responsible for the inhibitory effect of AEDKC, we investigated the effect of AEDKC on TNF-α/IFN-γ-induced activation of STAT1 and NF-κB. Previous studies have reported that the STAT and NF-κB signaling pathways contribute to the production of pro-inflammatory cytokines (TNF-α, IL-1β and IL-6) and chemokine (CCL17) in TNF-α/IFN-γ-induced HaCaT cells (29,30). As shown in Fig. 5, activation of STAT1 and NF-κB was reduced by AEDKC (1,000 µg/ml).