Six-week-old male BALB/c mice were obtained from Chongqing Medical University Animal House Center and maintained at Children’s Hospital of Chongqing Medical University Animal Care Centre. Animals were housed in plastic cages in a temperature-controlled (21°C) colony room on a 12/12 h light/dark cycle. Food and water were available ad libitum. All experiments and procedures were approved by Chongqing Medical University Animal Care and Use Committee. All efforts were made to minimize the number of animals used.

Fluoxetine and DNCB were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA). DNCB was dissolved in acetone-olive oil (4:1, v/v) and fluoxetine was dissolved in 0.9% sterile saline at a concentration of 10 mg/ml. The drug application and experimental tests were performed in a double-blinded manner.

Induction of AD-like skin lesions procedure is described in Figure 1 as described previously (Kim S. R. et al., 2014). The backs of mice were shaved with an electric clipper and depilatory cream a day before DNCB sensitization. The DNCB sensitization and challenge were performed for 5 weeks. For the sensitization process, a 1 cm × 1 cm gauze-attached patch was applied with 200 μl of 1% DNCB and attached to the shaved area twice per week. Two weeks after sensitization, the back skin was challenged with 200 μl of a 0.2% DNCB solution twice per week. This procedure was repeated for 3 weeks and fluoxetine was intraperitoneally administrated together. Behavioral tests were performed within 3 days after DNCB and fluoxetine treatments. At the end of the experiment, mice were sacrificed and samples were collected to evaluate the effects of DNCB and fluoxetine treatments (Figure 1).

The severity of DNCB-induced skin lesions was clinically assessed as previously described (Matsuda et al., 1997; Kim et al., 2008). Briefly, the total clinical severity score for skin lesions was calculated as the sum of the individual scores (0, no symptom; 1, mild; 2, moderate; 3, severe) for the following four AD signs and symptoms: erythema/hemorrhage, edema, excoriation/erosion, and scaling/dryness.

The animals were placed into an observation chamber for 10 min for acclimatization before the measurement of scratching behavior. The scratching behavior was defined as the hind limb scratches directed to the shaved area. The number of scratching behaviors was counted for 30 min each time.

Mice were placed in the center of a plus maze (each arm 30 cm) that was elevated 50 cm above the floor with two opposite open arms and two opposite closed arms (10-cm-tall walls on the closed arms) arranged at right angles. The number of entries and time spent in the closed and open arms were monitored for 10 min by ANY-Maze Video Tracking System. The maze was cleaned with 70% ethanol and water between tests.

Mice were placed in a cylinder of water (temperature 24–25°C; 10 cm in diameter, 30 cm in height) for 10 min. The depth of water was set to prevent animals from touching the bottom with their hind limbs. The latency to immobility, which was defined as floating or the least movement to maintain the head above the water, was recorded by ANY-Maze Video Tracking System (Stoelting Co.) from the side.

A portion of the skin biopsies were fixed in 4% paraformaldehyde. The skin sections (4 μm thick) were stained with hematoxylin and eosin (H&E) to evaluate the epidermal hyperplasia and the other sections were stained with toluidine blue to evaluate the infiltration of mast cells. Mast cells were counted in five parts of high-power fields (HPF) at 400× magnification.

Blood samples were collected by cardiac puncture under anesthesia, and sera were collected by centrifugation and stored at -80°C until use. Total IgE levels in plasma were determined by sandwich ELISA using the Sigma mouse IgE ELISA kit (St. Louis, MO, USA).

Total RNA was extracted from spleen tissue with the TRIzol reagent (Takara, Otsu, Shiga, Japan) according to the manufac turer’s instruction. The quantity and purity of total RNA were determined with a Nanodrop reader (Nanodrop Technologies, Wilmington, DE, USA). One microgram of total RNA was converted to the first-strand DNA with PrimeScript RT reagent Kit with gDNA Eraser (Takara, Otsu, Shiga, Japan) reverse transcriptase (Takara, Otsu, Shiga, Japan) and RNAsin (Takara, Otsu, Shiga, Japan). cDNA was amplified using gene-specific primers and SYBR^® Premix Ex Taq^TM II (Takara, Otsu, Shiga, Japan). Primer sequences were as follows: Interferon-γ (IFN-γ) (forward: 5′-CTCAAGTGGCATAGATGT-3′, reverse: 5′-GAGATAATCTGGCTCTGCAGGATT-3′); IL-2 (forward: 5′-CTCTACAGCGGAAGCACAGCA-3′, reverse: 5′-TGCCGCAGAGGTCCAAGTT-3′); IL-4 (forward: 5′-TGAACGAGGTCACAGGAGAAGG-3′, reverse: 5′-CACCTTGGAAGCCCTACAGACA-3′); IL-13 (forward: 5′-AGCATGGTATGGAGTGTGGACCTG-3′, reverse: 5′-CAGTTGCTTTGTGTAGCTGAGCAG-3′); β-actin (forward: 5′-GCACCACACCTTCTACAATGAGC-3′, reverse: 5′-GGATAGCAGCCTGGATAGCAAC-3′). The relative expression levels of target genes were normalized using β-actin as an internal control.

All data are expressed as the mean ± SEM. Statistical analysis of the results were performed by one-way analysis of variance (ANOVA) followed by Turkey’s test. Significance level was set at p < 0.05.

Previous studies have shown that paroxetine, a selective SSRI, is effective in the treatment of AD in both animal model and patients, with no apparent side-effects (Zylicz et al., 2003; Jiang et al., 2007; Stander et al., 2009). To determine whether fluoxetine is effective in the treatment of AD-like skin lesions, we chronically treated AD mice with fluoxetine (10 mg/kg, i.p.; Figure 1). Consistent with previous reports (Kim H. et al., 2014; Kim S. R. et al., 2014; Yoon et al., 2015), AD-like lesions developed following repeated application of DNCB to the back skin (Figure 2A). More importantly, chronic fluoxetine administration dramatically alleviated skin lesions (Figure 2A). To further compare the skin lesions of the mice with those of human AD (Mihm et al., 1976; Piloto Valdes et al., 1990), clinical severity of the dermatitis was scored as previously described (Leung et al., 1990). The result showed that DNCB treatment significantly increased the skin lesion score compared with normal control, whereas the skin lesion score was significantly decreased after fluoxetine treatment (control: n = 10; DNCB: n = 10; DNCB+fluoxetine: n = 14; Figure 2B). In addition, chronic fluoxetine treatment significantly suppressed DNCB-induced scratching behaviors (control: n = 10; DNCB: n = 10; DNCB+fluoxetine: n = 14; Figure 2C). Taken together, these results suggest that chronic fluoxetine treatment alleviates the DNCB-induced skin lesions.

A growing body of evidence has shown that AD in both adolescence and adulthood increases the risk of developing major depression and anxiety disorders in later life (Wahlgren, 1992; Ginsburg et al., 1993; Hashiro and Okumura, 1997; Cheng et al., 2015), and fluoxetine has been widely used as an antidepressant agent in clinic. Thus, we next introduced different behavioral tests, including forced swimming and elevated plus maze, to determine whether the alleviating effect of fluoxetine on AD is associated with its antipsychotic effect. Compared to normal controls, AD mice displayed apparent depressive-like symptom, as reflected by much shorter latency to immobility during forced swimming test (control: n = 10, 186.8 ± 5.0 s; DNCB: n = 10, 157.2 ± 10.4 s, p < 0.05 vs. control; Figure 3A). In agreement with an anti-depressive effect of fluoxetine, it succeeded in abolishing the influence of DNCB on depressive behavior (DNCB+fluoxetine: n = 14, 228.3 ± 15.2 s, p < 0.05 vs. control, p < 0.01 vs. DNCB; Figure 3A).

Next, using elevated plus maze paradigm, we tested the effect of fluoxetine on anxiety-like behavior in AD. The results showed that both the number of entry into open arms (control: n = 10, 25.5 ± 3.9; DNCB: n = 10, 11.1 ± 1.5, p < 0.05 vs. control; Figure 3B) and the time spent in open arms (control: n = 10, 40.2 ± 5.5 s; DNCB: n = 10, 11.1 ± 3.1 s, p < 0.05 vs. control; Figure 3C) were significantly reduced in DNCB-treated mice during elevated plus maze test. Similar to forced swimming test, chronic fluoxetine treatment was able to restore the number of entry into the open arms (DNCB+fluoxetine: n = 14, 34.9 ± 4.9, p < 0.05 vs. control, p < 0.01 vs. DNCB; Figure 3B) and the time in open arms (DNCB+fluoxetine: n = 14, 61.7 ± 10.8 s, p < 0.05 vs. control, p < 0.01 vs. DNCB; Figure 3C) to control levels, indicating chronic fluoxetine administration reduces anxiety during AD development.

It has been widely accepted that AD is a common chronic inflammatory skin disease. We next wanted to determine the alterations of tissue inflammation and the accumulation of mast cells in AD-like skin lesions. Consistent with previous reports (Kim S. R. et al., 2014; Yoon et al., 2015), histopathological examination showed that the skins of the DNCB-treated group showed markedly epidermal hyperplasia as compared to the normal group, whereas this hyperplasia was significantly reduced by chronic fluoxetine treatment (control: n = 10, 17.0 ± 1.8 μm; DNCB: n = 10, 83.4 ± 13.2 μm, p < 0.01 vs. control; DNCB+fluoxetine: n = 14, 54.9 ± 5.8 μm, p < 0.01 vs. control, p < 0.05 vs. DNCB; Figures 4A,B). Similarly, the number of mast cells in the dermis of DNCB-treated mice was markedly increased as compared to the normal group, whereas this increment was significantly suppressed by chronic fluoxetine treatment (control: n = 10, 8.5 ± 1.7; DNCB: n = 10, 88.0 ± 8.5, p < 0.01 vs. control; DNCB+fluoxetine: n = 14, 57.3 ± 7.8, p < 0.01 vs. control, p < 0.05 vs. DNCB; Figures 4C,D).

Since serum IgE level is correlated with the severity of AD (Arshad and Holgate, 2001; Belloni et al., 2008; Ibler and Jemec, 2015; Montes-Torres et al., 2015), and the level of IgE is associated with Th2 or Th1 immunity (Snapper and Paul, 1987), we examined the serum level of IgE to evaluate the effect of fluoxetine on systemic Th1 and Th2 immunities in DNCB-treated mice. As shown in Figure 5, the serum level of IgE was markedly increased in the DNCB-treated group compared to normal control (control: n = 4, 11.1 ± 1.9 μg/ml; DNCB: n = 6, 19.6 ± 2.8 μg/ml, p < 0.01 vs. control; Figure 5). Chronic fluoxetine treatment restored the serum IgE to normal control level (DNCB+fluoxetine: n = 5, 11.6 ± 0.4 μg/ml, p > 0.05 vs. control, p < 0.05 vs. DNCB; Figure 5).

In the past decades, SSRIs including fluoxetine have been shown to exert several immunological effects, such as alteration of cytokine secretion (Janssen et al., 2010; Gobin et al., 2013, 2014). To evaluate the effects of fluoxetine on Th1 and Th2 immunities in AD, we examined the expression of IL-2, IFN-γ, IL-4, and IL-13 mRNAs in the spleen of experimental mice. The results showed that the IL-4 (control: n = 4, 100.0 ± 14.7; DNCB: n = 4, 288.6 ± 7.1, p < 0.01 vs. control; Figure 6A) and IL-13 (control: n = 6, 100.0 ± 6.5; DNCB: n = 5, 212.5 ± 36.1, p < 0.05 vs. control; Figure 6B) mRNA expressions were dramatically increased, whereas the IL-2 (control: n = 6, 100.0 ± 18.0; DNCB: n = 6, 84.0 ± 11.5, p > 0.05 vs. control; DNCB+fluoxetine: n = 5, 99.1 ± 12.6, p > 0.05 control, p > 0.05 vs. DNCB; Figure 6C) and IFN-γ (control: n = 6, 100.0 ± 17.6; DNCB: n = 6, 126.8 ± 4.4, p > 0.05 vs. control; DNCB+fluoxetine: n = 6, 102.6 ± 12.9, p > 0.05 control, p > 0.05 vs. DNCB; Figure 6D) mRNA expressions remained unchanged in DNCB-treated mice as compared to the normal group. More importantly, chronic fluoxetine treatment suppressed the DNCB-induced increase of IL-4 (DNCB+fluoxetine: n = 5, 145.2 ± 42.3, p > 0.05 control, p < 0.05 vs. DNCB; Figure 6A) and IL-13 mRNA expressions (DNCB+fluoxetine: n = 6, 132.1 ± 27.3, p > 0.05 control, p < 0.05 vs. DNCB; Figure 6B)