Use of Physcion to Improve Atopic Dermatitis-Like Skin Lesions through Blocking of Thymic Stromal Lymphopoietin

Physcion is well known for the treatment of carcinoma. However, the therapeutic effect of physcion on atopic dermatitis (AD) through the inhibition of thymic stromal lymphopoietin (TSLP) level remains largely unknown. In this study, we investigated the anti-AD effect of physcion using HMC-1 cells, splenocytes, and a murine model. Treatment with physcion decreased production and mRNA expression levels of TSLP, IL-6, TNF-α, and IL-1β in activated HMC-1 cells. Physcion reduced the expression levels of RIP2/caspase-1 and phospho (p)ERK/pJNK/pp38 in activated HMC-1 cells. Physcion suppressed the expression levels of pIKKβ/NF-κB/pIkBα in activated HMC-1 cells. Moreover, physcion attenuated the production levels of TSLP, IL-4, IL-6, TNF-α, and IFN-γ from activated splenocytes. Oral administration of physcion improved the severity of 2,4-dinitrochlorobenzene-induced AD-like lesional skin through reducing infiltration of inflammatory cells and mast cells, and the protein and mRNA levels of TSLP, IL-4, and IL-6 in the lesional skin tissues. Physcion attenuated histamine, IgE, TSLP, IL-4, IL-6, and TNF-α levels in serum. In addition, physcion inhibited caspase-1 activation in the lesional skin tissues. These findings indicate that physcion could ameliorate AD-like skin lesions by inhibiting TSLP levels via caspase-1/MAPKs/NF-kB signalings, which would provide experimental evidence of the therapeutic potential of physcion for AD.


Introduction
Atopic dermatitis (AD) is one of the most common relapsing inflammatory skin diseases in the world [1]. AD has an estimated prevalence of 15-20% in children and 2.6-8% in adults worldwide [2,3]. AD worsens quality of life and results in considerable burdens such as insufficient sleep, school and/or work absenteeism, psychological stress, and high medical costs [4].

Physcion Attenuates TSLP Level in PMA Plus Calcium Ionophore (PMACI)-Stimulated HMC-1 Cells
First, we determined the concentration of physcion that did not affect the viability of HMC-1 cells. Figure 2A shows that 2500 ng/mL of physcion had significant cytotoxicity to PMACI-stimulated HMC-1 cells (p < 0.05). Thus, we determined physcion concentrations at 2.5, 25, and 250 ng/mL in this study. In our previous study, we reported that TSLP secreted from mast cells played a pivotal role in the pathogenesis of AD [10]. The TSLP secreted from mast cells is regulated by intracellular calcium [21]. Thus, we clarified whether physcion could regulate TSLP levels by reducing the intracellular calcium in mast cells. As shown in Figure 2B, the stimulation with PMACI markedly increased intracellular calcium levels. However, the treatment with physcion suppressed the intracellular calcium levels ( Figure 2B). The beneficial effect on the calcium level of physcion (250 ng/mL)-treated group was similar to that of 2-bis(2-aminophenoxy)ethane-N,N,N0,N0-tetraacetic acid acetoxymethyl ester (BAPTA-AM, calcium chelator)-treated group. Next, we examined the regulatory effect of physcion on TSLP levels secreted from the PMACI-stimulated HMC-1 cells. The treatment with physcion (2.5, 25, and 250 ng/mL) significantly induced dose-dependent reductions

Physcion Attenuates TSLP Level in PMA Plus Calcium Ionophore (PMACI)-Stimulated HMC-1 Cells
First, we determined the concentration of physcion that did not affect the viability of HMC-1 cells. Figure 2A shows that 2500 ng/mL of physcion had significant cytotoxicity to PMACI-stimulated HMC-1 cells (p < 0.05). Thus, we determined physcion concentrations at 2.5, 25, and 250 ng/mL in this study. In our previous study, we reported that TSLP secreted from mast cells played a pivotal role in the pathogenesis of AD [10]. The TSLP secreted from mast cells is regulated by intracellular calcium [21]. Thus, we clarified whether physcion could regulate TSLP levels by reducing the intracellular calcium in mast cells. As shown in Figure 2B, the stimulation with PMACI markedly increased intracellular calcium levels. However, the treatment with physcion suppressed the intracellular calcium levels ( Figure 2B). The beneficial effect on the calcium level of physcion (250 ng/mL)-treated group was similar to that of 2-bis(2-aminophenoxy)ethane-N,N,N0,N0-tetraacetic acid acetoxymethyl ester (BAPTA-AM, calcium chelator)-treated group. Next, we examined the regulatory effect of physcion on TSLP levels secreted from the PMACI-stimulated HMC-1 cells. The treatment with physcion (2.5, 25, and 250 ng/mL) significantly induced dose-dependent reductions in the production and mRNA expression levels of TSLP in the PMACI-stimulated HMC-1 cells ( Figure 2C,G, p < 0.05). In addition, physcion (25 and 250 ng/mL) significantly suppressed the production and mRNA expression levels of IL-6, TNF-α, and IL-1β ( Figure 2D-F,H-J, p < 0.05). Physcion alone did not affect these levels in non-stimulated HMC-1 cells (Figure 2). Dexamethasone (DEX) also had a regulatory effect on the TSLP, IL-6, TNF-α, and IL-1β levels in activated HMC-1 cells ( Figure 2C-J, p < 0.05).

Physcion Downregulates RIP2 and Caspase-1 Expressions in PMACI-Stimulated HMC-1 Cells
In our previous reports, RIP2 and caspase-1 regulated TSLP levels in HMC-1 cells [15,21]. Thus, we examined the expression levels of RIP2 and caspase-1 in the PMACI-stimulated HMC-1 cells to clarify the underlying mechanisms of the regulatory effect of physcion on TSLP levels. As shown in Figure 3A,B, PMACI stimulation significantly increased the expression levels of RIP2 and caspase-1 (p < 0.05) however, the treatment with physcion (25 and 250 ng/mL) significantly decreased the expression levels of RIP2 and caspase-1 (p < 0.05). In addition, physcion (25 and 250 ng/mL) significantly suppressed caspase-1 activities in the PMACI-stimulated HMC-1 cells ( Figure 3C, p < 0.05). The regulatory effects of physcion (250 ng/mL) on expression levels of RIP2 and caspase-1 were similar to those of the DEX-treated group (Figure 3). with Physcion or Dexamethasone (DEX) for 1 h and then stimulated with PMACI for 7 h. Cell viabilities were analyzed with a MTT assay. (B) Physcion or BAPTA-AM was treated in Fura-2/AM-pretreated HMC-1 cells for 20 min. The HMC-1 cells were then stimulated with PMACI. Blank, non-stimulated cells; PMACI, PMACI-stimulated cells. HMC-1 cells were treated with Physcion or DEX for 1 h and then stimulated with PMACI for 7 h for ELISA or 5 h for polymerase chain reaction (PCR). The production levels of (C) TSLP, (D) IL-6, (E) TNF-α, and (F) IL-1β were detected by ELISA. The mRNA expression levels of (G) TSLP, (H) IL-6, (I) TNF-α, and (J) IL-1β were detected by Quantitative real-time PCR. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA expression levels were analyzed as a housekeeping gene for normalization. Data is expressed as fold induction relative to a vehicle group. #p < 0.05 vs. a non-stimulated group. *p < 0.05 vs. a PMACI-stimulated group.

Physcion Downregulates RIP2 and Caspase-1 Expressions in PMACI-Stimulated HMC-1 Cells
In our previous reports, RIP2 and caspase-1 regulated TSLP levels in HMC-1 cells [15,21]. Thus, we examined the expression levels of RIP2 and caspase-1 in the PMACI-stimulated HMC-1 cells to clarify the underlying mechanisms of the regulatory effect of physcion on TSLP levels. As shown in Figure 3A,B, PMACI stimulation significantly increased the expression levels of RIP2 and caspase-1 (p < 0.05) however, the treatment with physcion (25 and 250 ng/mL) significantly decreased the expression levels of RIP2 and caspase-1 (p < 0.05). In addition, physcion (25 and 250 ng/mL) significantly suppressed caspase-1 activities in the PMACI-stimulated HMC-1 cells ( Figure 3C, p < 0.05). The regulatory effects of physcion (250 ng/mL) on expression levels of RIP2 and caspase-1 were similar to those of the DEX-treated group ( Figure 3). Caspase-1 activity was determined with a caspase-1 assay kit. #p < 0.05 vs a non-stimulated group. *p < 0.05 vs a PMACI-stimulated group.

Physcion Relieves Pathological Changes of AD-Like Lesional Skin
We found that physcion inhibits TSLP levels as well as inflammatory cytokines levels in activated mast cells. Thus, we next wished to validate whether physcion could relieve AD-like symptoms using a DNFB-induced AD-like murine model. As shown in Figure 5A, repeated topical application of DNFB to the dorsal surface induced AD-like symptoms including erosion, erythema, scaling, and excoriation of lesional skin, however physcion markedly ameliorated these symptoms compared to those of the DNFB control group. Physcion also induced decreases in the thickness of the epidermis and infiltrations of inflammatory cells and mast cells in the lesional skin ( Figure 5B-D, p < 0.05). DEX also ameliorated the pathological changes of the lesional skin similar to those observed in the physcion-treated group ( Figure 5).

Physcion Relieves Pathological Changes of AD-Like Lesional Skin
We found that physcion inhibits TSLP levels as well as inflammatory cytokines levels in activated mast cells. Thus, we next wished to validate whether physcion could relieve AD-like symptoms using a DNFB-induced AD-like murine model. As shown in Figure 5A, repeated topical application of DNFB to the dorsal surface induced AD-like symptoms including erosion, erythema, scaling, and excoriation of lesional skin, however physcion markedly ameliorated these symptoms compared to those of the DNFB control group. Physcion also induced decreases in the thickness of the epidermis and infiltrations of inflammatory cells and mast cells in the lesional skin ( Figure 5B-D, p < 0.05). DEX also ameliorated the pathological changes of the lesional skin similar to those observed in the physcion-treated group ( Figure 5).

Physcion Reduces TSLP, IL-4, And IL-6 Expression Levels in Lesional Skin of DNFB-Induced AD-Like Murine Model
Then, we further examined the regulatory effects of physcion on TSLP, IL-4, and IL-6 expression levels in the lesional skin tissues. The protein levels of TSLP, IL-4, and IL-6 were significantly reduced by the treatment with physcion ( Figure 7A-C, p < 0.05). Physcion also significantly decreased the mRNA expression levels of TSLP, IL-4, and IL-6 in the lesional skin tissues ( Figure 7D

Physcion Reduces TSLP, IL-4, And IL-6 Expression Levels in Lesional Skin of DNFB-Induced AD-Like Murine Model
Then, we further examined the regulatory effects of physcion on TSLP, IL-4, and IL-6 expression levels in the lesional skin tissues. The protein levels of TSLP, IL-4, and IL-6 were significantly reduced by the treatment with physcion ( Figure 7A-C, p < 0.05). Physcion also significantly decreased the mRNA expression levels of TSLP, IL-4, and IL-6 in the lesional skin tissues ( Figure 7D-G, p < 0.05). The regulatory effects of physcion on TSLP, IL-4, and IL-6 levels in the lesional skin tissues were similar to those of DEX-treated group (Figure 7).

Physcion Reduces TSLP, IL-4, And IL-6 Expression Levels in Lesional Skin of DNFB-Induced AD-Like Murine Model
Then, we further examined the regulatory effects of physcion on TSLP, IL-4, and IL-6 expression levels in the lesional skin tissues. The protein levels of TSLP, IL-4, and IL-6 were significantly reduced by the treatment with physcion ( Figure 7A-C, p < 0.05). Physcion also significantly decreased the mRNA expression levels of TSLP, IL-4, and IL-6 in the lesional skin tissues ( Figure 7D-G, p < 0.05). The regulatory effects of physcion on TSLP, IL-4, and IL-6 levels in the lesional skin tissues were similar to those of DEX-treated group (Figure 7).

Physcion Suppresses Caspase-1 Activation in AD-Like Lesional Skin
We found that physcion inhibited TSLP levels via caspase-1 signaling pathways in activated HMC-1 cells. Thus, to further understand the roles of physcion on caspase-1 signaling during the development of AD, we finally examined a regulatory effect of physcion on the activity and protein expression level of caspase-1 in the lesional skin tissues. Figure 8A shows that the activity of caspase-1 was enhanced in the lesional skin tissues from AD-like mice, which was repressed by physcion (p < 0.05). Physcion also significantly reduced the protein expression level of caspase-1 in the lesional skin tissues ( Figure 8B,C, p < 0.05). The regulatory effects of physcion on caspase-1 activation in the lesional skin tissues were similar to those of DEX-treated group (Figure 8).
HMC-1 cells. Thus, to further understand the roles of physcion on caspase-1 signaling during the development of AD, we finally examined a regulatory effect of physcion on the activity and protein expression level of caspase-1 in the lesional skin tissues. Figure 8A shows that the activity of caspase-1 was enhanced in the lesional skin tissues from AD-like mice, which was repressed by physcion (p < 0.05). Physcion also significantly reduced the protein expression level of caspase-1 in the lesional skin tissues ( Figure 8B,C, p < 0.05). The regulatory effects of physcion on caspase-1 activation in the lesional skin tissues were similar to those of DEX-treated group (Figure 8).

Discussion
The crosslinking of FcεRI-bound IgE with multivalent antigen initiates activation of mast cells, then mast cell activation results in protein kinase C (PKC) activation as well as intracellular calcium elevation [29]. To produce similar conditions, we selected PMA to activate PKC and calcium ionophore to increase intracellular calcium level in this study. TSLP mRNA expression and production increased by stimulation with PMACI [15]. Increased TSLP expression was shown in skin lesions from patients with AD [30]. Furthermore, increased TSLP levels resulted in an exacerbation of scratching behavior in the AD model [31]. Deletion of TSLP improved AD-like skin lesions in mice [32]. Pretreatment with physcion decreased TSLP levels in HMC-1 cells, splenocytes, serum as well as skin lesions (Figure 2, Figure 6, Figure 7 and Table 1). Thus, we presume that physcion may be beneficial to prevent and/or treat atopic diseases. Patients with AD possess higher levels of proinflammatory cytokines such as IL-6, TNF-α, IL-1β, IL-4, and IFN-γ [33][34][35]. Our results showed that physcion suppresses the levels of IL-6, TNF-α, IL-1β, IL-4, and IFN-γ, suggesting the potential of physcion in the treatment of AD.
Han et al. [21] reported that calcium chelator decreases RIP2 levels in HMC-1 cells, indicating that RIP2 is a downstream factor of calcium. In addition, Humke et al. [36] reported that RIP2 activates caspase-1. An increment of intracellular calcium was inhibited by pretreatment with physcion in HMC-1 cells (Figure 2). Thus, we assume that physcion may decrease TSLP levels via inhibiting of calcium/RIP2/caspase-1 signal cascade in HMC-1 cells.
Caspase-1 is activated by proinflammatory stimuli [36]. Lots of research has shown that caspase-1 is activated by proinflammatory stimuli such as PMACI [37][38][39]. Caspase-1 overexpression leads to AD-like skin lesions in mice [40]. Downregulation of caspase-1 by inhibitor treatment ameliorates AD symptoms in mice [41]. Our results showed that physcion downregulates caspase-1 activation in HMC-1 cells as well as AD-like skin lesions (Figures 3 and 8), presenting that physcion may ameliorate AD symptoms by the downregulation of caspase-1 activation.
The knock out of RIP2 shows decreased phosphorylation levels of ERK, JNK, and p38 MAPKs [42]. Treatment with caspase-1 inhibitor attenuates phosphorylation of p38, which suggests that caspase-1 is an upstream factor of p38 [43]. Furthermore, caspase-1 inhibition suppresses phosphorylation of ERK, JNK, and p38 MAPKs in human macrophages [44]. The results of the present study present that physcion suppresses the phosphorylation of ERK, JNK, and p38 in HMC-1 cells (Figure 4). Thus, we presume that suppression of TSLP by pretreatment with physcion, at least in part, may be mediated by RIP2/caspase-1/MAPKs signaling.
Lee and Ziegler [45] suggested that inducible expression of TSLP is mediated by NF-κB. Moon and Kim [15] also suggested that TSLP production is controlled by NF-κB in HMC-1 cells. Pretreatment with physcion suppressed phosphorylation of IKKβ and IκBα as well as activation of NF-κB (Figure 4), suggesting that physcion would suppress TSLP via inhibiting of NF-κB.
Histological AD features in human are epidermal thickening and inflammatory infiltrate [46]. Physcion ameliorates AD-like skin lesions, epidermal thickening, and inflammatory cell infiltration ( Figure 5). Furthermore, mast cell infiltration is upregulated in the skin lesions of AD, which suggests that mast cells play a role in AD [47]. Mast cell infiltration into skin lesions was downregulated by the oral administration of physcion ( Figure 5). Patients with AD showed higher histamine levels in serum compared with those in healthy subjects [48]. In addition, remarkable amelioration in the symptoms of AD resulted from anti-histamine treatment [48]. Our results showed that physcion reduces serum histamine levels and scratching behaviors ( Figure 6). Therefore, we assume that physcion may be used as an anti-histamine therapy for AD. In this study, figures and tables use different concentrations of physcion, some use ng/mL, others µg/kg. While it is similar in biochemical experiments, it is not the same in case of animal experiments. Nevertheless, a lot of researchers use the same method in various models [49][50][51][52]. Thus, we used ng/mL or µg/kg to show the concentrations of physcion.
In conclusion, we showed that treatment with physcion decreases mRNA expression and production of TSLP, IL-6, TNF-α, and IL-1β in activated HMC-1 cells. Pretreatment with physcion downregulated the levels of intracellular calcium, activation of RIP2 and caspase-1, phosphorylation of p38, JNK, and ERK, activation of NF-κB, as well as phosphorylation of IKKβ and IκBα in activated HMC-1 cells. In addition, physcion reduced the production of TSLP, TNF-α, IL-4, IL-6, and IFN-γ from activated splenocytes. Oral administration of physcion improved AD-like skin lesions and reduced the levels of TSLP, IL-4, and IL-6, as well as caspase-1 activation in the skin lesions. In serum, histamine, IgE, TSLP, TNF-α, IL-6, and IL-4 levels were downregulated by physcion. These findings indicate that physcion could ameliorate AD-like skin lesions by inhibiting TSLP levels via caspase-1/MAPKs/NF-kB signalings, which would provide experimental evidence of the therapeutic potential of physcion for the treatment of AD.

Physcion Preparation
Physcion was dissolved in dimethyl sulfoxide (DMSO) according to a report by Shen et al. [53] and diluted with distilled water. The doses were decided according to a report by Shen et al. [53]. DEX was prepared according to a report by Chen et al. [54].

MTT Assay
An MTT assay was performed to evaluate cell viability. HMC-1 cells (3 × 10 4 cells/well) were seeded into a 24-well plate and were then treated with different concentrations of physcion for 1 h. The cells were stimulated with PMACI for 7 h. 50 µL of MTT (5 mg/mL) solution was added to 500 µL of cell culture medium at 37 • C for 4 h. The crystallized formazan was then dissolved in DMSO. Optical density (O.D.) was detected at 540 nm by an enzyme-linked immunosorbent assay (ELISA) reader (Versa Max, Molecular Devices, Sunnyvale, CA, USA).

Intracellular Calcium Levels
HMC-1 cells (1 × 10 5 ) were pretreated with Fura-2/AM (4 µM) in IMDM supplemented with 10% heat-inactivated FBS for 30 min. After the cells were washed with a calcium free medium containing EGTA (0.5 mM), the cells were treated with physcion or BAPTA-AM (10 µM) and stimulated with PMACI. The kinetics of intracellular calcium was measured for 100 s with a spectrofluorometer (excitation 360 nm, emission 450 nm, Thermo Fisher Scientific Inc., Shanghai, China).

Cytokines Assay
The levels of TSLP, IL-6, TNF-α, IL-1β, IL-4, IFN-γ, and IgE from HMC-1 cells supernatant, splenocytes supernatant, serum, or lesional skin homogenates were detected by ELISA according to the manufacturer's instructions (R & D system Inc. and Pharmingen). The protein concentration was analyzed using a bicinchoninic acid protein assay kit.

Quantitative Real-Time Polymerase Chain Reaction (PCR) And Quantitative Reverse-Transcription PCR
Total RNA from HMC-1 cells or lesional skin tissues was isolated with an easy-BLUE™ RNA extraction kit (iNtRON Biotech, Sungnam, Korea). A cDNA synthesis kit (Bioneer Corporation, Daejeon, Korea) was used for synthesizing first-strand cDNA from total RNA. Real-time PCR for HMC-1 cells was performed using a SYBR Green master mix with an ABI StepOne real time-PCR System (Applied Biosystems, Foster City, CA, USA). Reverse-transcription PCR for lesional skin tissues was performed using an i-MAX™ II DNA polymerase kit (iNtRON Biotech, Sungnam, Korea) with a C1000 Touch Thermal Cycler (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Primer sequences were as follows: hTSLP (5 TAT GAG TGG GAC CAA AAG TAC CG 3  Reverse-transcription PCR products were electrophoresed on a 1.5% ethidium bromide agarose gel. Each mRNA level was quantified using the Image J software (National Institute of Health, Bethesda, MD, USA).

Western Blotting
The HMC-1 cells lysed with RIPA buffer to prepare whole cell extracts and lesional skin homogenates were separated on 12% SDS-PAGE and transferred to nitrocellulose membranes. Equivalent amounts of protein were blocked with phosphate buffered saline with Tween 20 containing 5% non-fat dry milk and probed with primary antibodies against pERK, ERK, pJNK, JNK, pp38, p38, RIP2, caspase-1, pIKKβ, NF-κB, pIκBα, PARP, and GAPDH at 4 • C overnight. The membrane was then probed with peroxidase-conjugated second antibodies. Immunodetection was carried out using an enhanced chemiluminescence solution (DoGenBio Co., Seoul, Korea). Each band intensity was quantified using the Image J software (National Institute of Health).

Caspase-1 Activity Assay
The enzymatic activities of caspase-1 in HMC-1 cells lysates and lesional skin homogenates were analyzed with a caspase-1 assay kit according to the manufacturer's instructions (R & D system Inc.).

Nuclear Extracts And Cytoplasmic Extracts
The pellet from HMC-1 cell was resuspended in 40 µL of cytoplasmic extract buffer (0.1 mM EDTA, 10 mM HEPES/KOH, 2 mM MgCl2, 1 mM dithiothreitol, 10 mM KCl, and 0.5 mM phenylmethylsulfonyl fluoride) for 15 min on ice and then lysed with 0.6 µL of 10% Nonidet P-40. After centrifuge, the cytoplasmic extract was removed from the pellet to a clean tube. The nuclei pellet was washed in 20 µL of cytoplasmic extract buffer and lysed with 40 µL of nuclear extract buffer (50 mM HEPES/KOH, 0.1 mM EDTA, 1 mM dithiothreitol, 50 mM KCl, 300 mM NaCl, 10% glycerol, and 0.5 mM phenylmethylsulfonyl fluoride) for 20 min on ice. After centrifuge, supernatant (nuclear fraction) was transfered to a clean tube.

Animals
All experiments were performed according to internationally-accepted standards for laboratory animal use and care, as found in the United States guidelines (NIH publication no. 85-23, revised in 1985). Studies involving animals was approved from the animal care committee of Kyung Hee University [KHUASP (SE)-18-022]. Female BALB/c mice of 8 weeks (Dae-Han Experimental Animal Center) were raised under conventional conditions (22 ± 1 • C, 40%-50% relative humidity, and 12:12 h light/dark cycle).

DNFB-Induced AD-Like Lesional Skin
DNFB-induced AD-like lesional skin was prepared as described previously [6] ( Figure 1B). 100 µL of acetone or 0.15% DNFB was topically applied to shaved abdominal skin on 1st day. The shaved dorsal surface was applied with 50 µL of acetone as a vehicle group or 50 µL of DNFB as a control group on the eighth day and these applications were processed twice a week for 3 weeks. At the same time, 0.0025% DMSO (vehicle group), physcion (250 µg/kg), or DEX (10 nM) was orally administered to the mice three times a week for 3 weeks. For example, 200 µL of physcion solution (25 µg/mL) was orally administered to a mouse (20 g, body weight). Serum and lesional skin were collected 4 h after the last DNFB sensitization after anesthesia.

Histological Analysis
The lesional skin tissues fixed in 4% formalin were embedded in paraffin. Tissue slides (4 µm) were deparaffinized and rehydrated for staining with hematoxylin and eosin (H&E) or toluidine blue.

Histamine Assay
Serum histamine levels were analyzed according to o-phthaldialdehyde spectrofluorometric procedure, as previously described [55].

Statistics
The results are expressed as the mean ± standard error of mean (SEM). Statistical analysis was conducted using the IBM SPSS v23 statistics software (Armonk, NY, USA). An independent t-test was conducted for comparing differences between two groups (Blank group vs. PMACI control group or vehicle group vs. DNFB control group). ANOVA with Tukey post-hoc test was conducted for comparing differences among the PMACI control group/DNFB control group vs. the physcion or DEX-treated group. A p value <0.05 was considered significant.