Twenty-eight, four to five-week-old female NC/Nga mice were purchased from Charles River Japan (Yokohama, Japan). The mice were housed and bred under conventional conditions at the animal laboratory of the Uijeongbu St. Mary's Hospital. All animal procedures were approved by the Ethics of Animal Experimentation Committee of Uijeongbu St. Mary's Hospital, and conformed to international standards.

Dfb ointment was prepared by Biostir Inc. (Kobe, Japan). One gram of Dfb ointment contained 41.7 mg of protein, 58.5 µg of Der f 1, and 22.2 µg of Der f 2.

After 1 week of adaptation, the mice were randomly allocated into 4 groups: Group 1, the untreated negative control (CONT); Group 2, Dfb ointment-induced AD as a positive control (AD); Group 3, Dfb ointment+distilled water bathing (AD+DW); and Group 4, Dfb ointment+mineral water bathing (AD+MW). In the first induction, the dorsal hair of the mice was clipped using an electric shaver, and the residual hair was depilated using a hair removal cream. Groups 2 to 4 were challenged by the topical application of 100 mg of Dfb ointment on the shaved dorsal skin and the surface of both ears, while untreated Group 1 served as a negative control. In the second induction, barrier disruption was achieved by 150 µl of 4% sodium dodecyl sulfate treatment on the shaved dorsal skin and the surface of both ears 3 hours before Dfb ointment application. These procedures were repeated 3 to 4 times a week for up to 2 weeks. The experiment schedule is summarized in Fig. 1.

After 2 weeks of Dfb application, when the phenotype of AD-like chronic allergic dermatitis had been established, each of Groups 3 and 4 were immersed in distilled or mineral water baths for 5 minutes (Fig. 2). The temperature of the water was 37℃ to 39℃. Bathing was performed daily for 2 weeks. Group 2 mice were untreated. Suanbo Mineral Water has been classified in the category of waters with low mineral content (total dissolved solids: 348 mg/L), and sodium bicarbonate is the major component. The mean water temperature was 42.40℃, and the pH was 8.39. Compared to other Korean hot springs, Suanbo is richer in its proportions of calcium and sulfate (Table 1). Transepidermal water loss (TEWL) was measured on individual flanks with a skin-evaporative water recorder (Tewameter MPA5®; Courage & Khazawa, Köln, Germany) immediately before bathing at the baseline, on experimental days 1, 3, 4, 7, and 11, and 12 hours after the final bathing. Assessments were carried out in an air-conditioned animal room (20℃ to 21℃; relative humidity 27%).

The severity of dermatitis was grossly assessed by the scoring AD (modified SCORAD) method before bathing at the baseline, on experimental days 3, 7 and 11, and 12 hours after the final bathing. The degree of each symptom was scored as 0 (absence), 1 (mild), 2 (moderate), and 3 (severe). This scoring was based on the severity of 1) erythema/hemorrhage, 2) scaling/dryness, 3) edema, and 4) excoriation/erosion. The sum of the individual scores (minimum 0 and maximum 12) was taken as the dermatitis score. Assessment was performed by an investigator who was blind to the grouping of the animals.

The sampling of blood, skin, and spleen was performed 12 hours after the final bathing. The dorsal skin was excised and divided into two specimens, one for histopathological evaluation and the other for mRNA extraction. Serum samples were stored at -70℃ until analysis. The spleen was processed for flow cytometric analysis.

The skin lesion samples were fixed in 10% formaldehyde, embedded in paraffin, and stained with standard H&E.

Paraformaldehyde-fixed cryosections were incubated overnight at 4℃ with antibodies against mouse anti-CD1 (sc-9161; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-CD4 (sc-70671; Santa Cruz Biotechnology), and anti-CD8 (H-160; Santa Cruz Biotechnology). Immunoreactivity was detected with a peroxidase-conjugated secondary antibody against mouse and rabbit immunoglobulin (Ig) G (DakoCytomation, Glostrup, Denmark), and subsequently counterstained with hematoxylin. Staining with secondary antibodies without primary antibodies was done to rule out non-specific antibody binding.

The serum levels of interferon (IFN)-γ, interleukin (IL)-4, IL-5, and IgE were measured using specific ELISA kits according to the manufacturer's instructions (Mouse IFN-γ: #430804, ELISA MAX™ Deluxe Sets; BioLegend, San Diego, CA, USA, Mouse IgE: #432404, ELISA MAX™ Deluxe Sets; BioLegend, IL-4: #555232, Mouse IL-4 ELISA Set; BD Bioscience, San Diego, CA, USA, IL-5: #555236, Mouse IL-5 ELISA Set; BD Bioscience).

To study the gene expression of the IFN-γ, IL-4, and IL-5 of dorsal skin, total RNA was extracted from the skin using TRIzol (Invitrogen, Carlsbad, CA, USA) in accordance with the manufacturer's instructions. Total RNA was transcribed into cDNA using a reverse transcription system (Promega, Madison, WI, USA). The primers used for the reverse transcription-polymerase chain reaction (PCR) were as follows: 5'-TTT GCA GCT CTT CCT CAT GG-3' (forward) and 5'-ACC ATC CTT TTG CCA GTT CC-3' (reverse) for IFN-γ, and 5'-ATA TCC ACG GAT GCG ACA AA -3' (forward) and 5'-AAG CCC GAA AGA GTC TCT GC-3' (reverse) for IL-4. For the detection of IL-5, 5'-GAC CTT GAC ACA GCT GTC CG-3' (forward) and 5'-AGC ATT TCC ACA GTA CCC CC-3' (reverse) primers were used. Glyceraldehyde 3-phosphate dehydrogenase genes were used as controls for RNA input. Real-time PCR with SYBR green detection was conducted using a Maxime PCR premix Kit (Intron Biotechnology, Seongnam, Korea) with an ABI PRISM 7000 sequence detection system (Applied Biosystems, Foster City, CA, USA) with 50 ng of cDNA.

To analyze Th1, spleen cells were stained with PE anti-mouse CD4 (100407; BioLegend). After cell fixation, pellet cells were washed three times in an 80:20 mixture of fluorescence-activated cell sorting buffer and permeabilization buffer, followed by the suspension of cells in the same buffer mixture with primary antibody for 30 minutes on ice. The permeabilization buffer contained 0.5% saponin and 1% bovine serum albumin in phosphate buffered saline. Subsequently, cells were stained with antigen-presenting cell (APC) anti-mouse IFN-γ (505809; BioLegend). For the analysis of Th2, cells were stained with PerCP anti-mouse CD4 (100431; BioLegend), followed by intracellular staining with Alexa Fluor 488 anti-mouse IL-4 (504111; BioLegend) and PE anti-mouse IL-9 (514103; BioLegend). For the T_reg cells, cells were fixed and permeabilized, then stained with Alexa Fluor 488 anti-mouse/rat/human Foxp-3 (320011; BioLegend). Isotype controls were used: PE-Rat IgG2b, κ (for PE anti-mouse CD4), PerCP-Rat IgG2b, κ (for PerCP anti-mouse CD4), APC-Rat IgG1, κ (for APC anti-mouse IFN-γ), Alexa Fluor 488-Rat IgG1, κ (for Alexa Fluor anti-mouse IL-4), PE-Rat IgG1, κ (for PE anti-mouse IL-9), and Alexa Fluor 488-Mouse IgG1, κ (for Alexa Fluor 488 anti-mouse/rat/human Foxp-3). First, the isotype-only tubes under the FL1-versus-FL2 dot-plot with quadrant division were measured. The compensation was adjusted until the cell population gathered into the lower-left quadrant. Data were acquired on a FACS Calibur system (BD Bioscience) and analyzed using CellQuest software (BD Bioscience).

All data are expressed as the mean±standard deviation. One-way analysis of variance followed by Tukey's multiple comparison test were used for the statistical analysis. In all cases, p<0.05 were considered significant.

After the repeated application of Dfb ointment for 2 weeks, acute erythematous and edematous dermatitis occurred first, followed by a persistent chronic dermatitis with severe erythema, hemorrhage, scarring, erosion, and excoriation (Fig. 3A). Two weeks after the first challenge, the skin severity scores of the AD group reached a score of more than 9 points (Fig. 3B). Severe scratching behavior of the back and ears with hind paws was also observed. However, in the CONT group, no obvious skin lesions were observed (Fig. 3A). Repeated Dfb application produced marked epidermal hyperplasia, acanthosis, hyperkeratosis, and parakeratosis, accompanied by a prominent inflammatory infiltration in the dermis (Fig. 4A). In parallel with the emergence of epidermal hyperplasia and inflammatory cell infiltration, serum levels of IgE following Dfb ointment application were significantly elevated compared with the CONT group (Fig. 5A). Such Dfb-induced features are representative of AD8.

The efficacy of bathing in mineral water was examined, in Dfb-AD mice, as assessed by changes in clinical appearance, histologic features, inflammatory infiltrates, and basal TEWL. Bathing in mineral water for 2 weeks alleviated the severity of the AD-like skin lesions (Fig. 3A) and markedly reduced dermatitis scores (Fig. 3B) and TEWL (Fig. 3C). Moreover, inflammatory infiltrates also declined after bathing in mineral water (Fig. 4A). The dermatitis scores, TEWL, and inflammatory cell infiltration also declined after bathing in distilled water compared with the AD group, but to a lesser extent than the AD+MW group. The AD group showed no remarkable changes in dermatitis scores, TEWL (Fig. 3B, C), and inflammatory cell infiltration (Fig. 4A).

In parallel with these apparent therapeutic benefits, bathing in mineral water significantly normalized Dfb-induced epidermal hyperplasia (Fig. 4A). In contrast, epidermal hyperplasia, hyperkeratosis, and parkeratosis were still apparent in the AD+DW group (Fig. 4), respectively, compared to the AD+MW group.

The effect of balneotherapy in the infiltration of CD1⁺, CD4⁺ and CD8⁺ cells was further investigated. As shown in Fig. 4B, bathing in mineral water decreased CD1⁺ cell infiltration. No significant reduction in CD1⁺ cells was found in the AD+DW group. The effect of balneotherapy in the infiltration of CD4⁺ and CD8⁺ cells was also investigated. As shown in Fig. 4C and D, the infiltration of CD4⁺ and CD8⁺ cells also remarkably declined after bathing in mineral water. Although the reduction of CD4⁺ and CD8⁺ cells in the AD+DW group was also observed, it was not to the same extent as in the AD+MW group.

The total serum IgE levels after 2 weeks of AD induction were markedly higher compared to the CONT group (Fig. 5A). The AD+MW and AD+DW groups failed to reduce total serum IgE levels following Dfb ointment application, and there was an increase in IL-4 levels following Dfb ointment application. Mineral water bathing seemed to suppress Dfb ointment-induced IL-4 levels, unlike distilled water (Fig. 5B). However, no statistically significant difference was found in IL-4 levels between these two groups. There was no significant difference in IFN-γ and IL-5 levels among the CONT and all experimental groups (Fig. 5C, D).

There was a significant increase in the mRNA expression of IFN-γ and IL-4 following AD induction (Fig. 6A, B). The Dfb ointment-induced mRNA expression of IL-4 was reduced by mineral water bathing, but it was not statistically significant (Fig. 6B). However, both the AD+MW and AD+DW groups failed to reduce Dfb ointment-induced IFN-γ mRNA expression (Fig. 6A). No significant difference was found in IL-5 among all the groups (Fig. 6C).

As shown in Fig. 7, mineral water bathing seemed to suppress Dfb ointment-induced Th2 cell proliferation. In contrast to the suppressive effect on Th2 cells, differentiation into T_reg cells was promoted with mineral water bathing compared to the AD+DW group. Neither mineral water nor distilled water treatment reduced Dfb ointment-induced Th1 cell proliferation.