Sphingomyelin maintains the cutaneous barrier via regulation of the STAT3 pathway

Epidermal tissues play vital roles in maintaining homeostasis and preventing the dysregulation of the cutaneous barrier. Sphingomyelin (SM), a sphingolipid synthesized by sphingomyelin synthase (SMS) 1 and 2, is involved in signal transduction via modulation of lipid‐raft functions. Though the implications of SMS on inflammatory diseases have been reported, its role in dermatitis has not been clarified. In this study, we investigated the role of SM in the cutaneous barrier using a dermatitis model established by employing Sgms1 and 2 deficient mice. SM deficiency impaired the cutaneous inflammation and upregulated signal transducer and activator of transcription 3 (STAT3) phosphorylation in epithelial tissues. Furthermore, using mouse embryonic fibroblast cells, the sensitivity of STAT3 to Interleukin‐6 stimulation was increased in Sgms‐deficient cells. Using tofacitinib, a clinical JAK inhibitor, the study showed that SM deficiency might participate in STAT3 phosphorylation via JAK activation. Overall, these results demonstrate that SM is essential for maintaining the cutaneous barrier via the STAT3 pathway, suggesting SM could be a potential therapeutic target for dermatitis treatment.


| INTRODUCTION
The epidermal tissue maintains the homeostasis in the cutaneous barrier and prevents the penetration of undesired substances, including several allergens, into the skin from the environment. 1 Excessive immunological reactions in response to the external aggressors play an important role in developing cutaneous inflammation, such as atopic dermatitis (AD), which suggests that cutaneous barrier dysfunction and excessive immune responses are highly relevant to each other.
Lipid-rafts, specific microdomains on the plasma membrane enriched in sphingomyelin (SM) and cholesterol, are thought to be important signaling platforms. 2,3 SM, one of the essential sphingolipids, constitutes nearly 85% of all sphingolipids and 10%-20% of all cellular membrane phospholipids. 4 SM is widely distributed in the myelin of neural tissues, lung surfactants, and the epidermis of the skin, in addition to lipid membranes in higher animals. Recently, several studies have shown that SM regulates signal transduction via modulation of lipid-raft functions. [5][6][7] The SM signal transduction pathway is induced by various cytokines such as tumor necrosis factor α (TNFα), interferon γ (IFNγ), and interleukin 1 (IL-1) family, and functions as a major player in the signaling pathways of apoptosis, cell differentiation, stress responses, inducers of cell damage, and cell cycle arrest.
Sphingomyelin synthase (SMS) synthesizes SM by transferring the phosphatidyl head group, phosphatidylcholine, onto the primary hydroxyl of ceramide. 6 SMS exists in two isoforms, SMS1 and SMS2 (encoded by SGMS1 and SGMS2, respectively), which have been cloned and characterized by their cellular localization. 8 They are abundant in the Golgi apparatus and plasma membrane, 9 and they are known to contribute to the maintenance of intracellular SM levels. 10,11 It has been shown that SMS is associated with inflammation, such as pneumonia and colitis 11,12 ; however, its role in dermatitis has not been elucidated.
To date, several investigations have revealed that ceramide, a common substrate of sphingolipids in lipid membranes, maintains the cutaneous barrier and regulates cutaneous inflammation [13][14][15] ; however, the involvement of other sphingolipids has not been elucidated. Recently, it has been reported that lipid rafts modulate the activation of the signal transducer and activator of the transcription 3 (STAT3) signaling pathway. 16 STAT3 is involved in multiple processes, including early development, cellular proliferation, survival, and differentiation, 17,18 and is also highly expressed in the skin of AD and psoriasis. 19,20 STAT3 is activated by many extracellular signaling molecules such as cytokines, growth factors, and hormones. 21 Moreover, the activation of STAT3 is accomplished by its phosphorylation on a tyrosine residue, leading to nuclear translocation and transcriptional regulation of target genes.
Based on this background information, we hypothesized that SM deficiency would impair the cutaneous barrier via the STAT3 pathway. Therefore, this study was aimed to investigate the effects of SMS deficiency on cutaneous inflammation, such as severe dermatitis, cutaneous barrier dysfunction, and epidermal thickness. Additionally, we also investigated its underlying mechanisms.

| Animal experiments
All animal experiments were performed according to the established guidelines of the "Law for the Care and Welfare of Animals in Japan" and approved by the Animal Experiment Committee of Azabu University (Approval No. 180316-5). Eight to 10 weeks old, C57BL/6N mice acquired from Charles River Laboratories (Kanagawa, Japan) were used for the study. Mice were housed in plastic cages in an air-conditioned room at 24°C under a 12 h light-dark cycle (light on at 7:00 a.m.) with food and water available ad libitum under Specific Pathogen Free conditions. At the end of the experiment, all animals were euthanized by cervical dislocation under 2% isoflurane anesthesia. Sgms1-and Sgms2-knock out (KO) mice were generated following the previously established protocols. 22,23

| A mouse model of dermatitis
A mouse model of dermatitis was generated based on a previously described method with minor modifications. 24 Briefly, at least 24 h before the first elicitation, hair of the rostral back region of all mice was depilated with an electric clipper and shaver. Barrier disruption was performed by 150 µl of 4% sodium dodecyl sulfate (Cat. No. 194-13985, Wako Pure Chemical Industries) treatment on the shaved dorsal skin 3 h before elicitation. Elicitation was performed by topical application of 50 mg Dermatophagoides farinae (Df) ointment (Biostir AD, Biostir Inc., Ohsaka, Japan) or ointment base (hydrophilic petrolatum) onto the shaved dorsal skin. The same procedure was repeated twice a week for two weeks. Crude sphingomyelin fraction in butter (BSM; GENUINE R&D, Fukuoka, Japan) based on hydrophilic petrolatum (contains 10% as sphingomyelin) was applied to the shaved dorsal skin every other day for one week, followed by Df application twice a week. The amount of SM applied to the dorsal skin of the SM-treated group was calculated based on the SM content in the C57BL/6N wild-type (WT) mice back skin. 25 The Toluene-2,4-diisocyanate (TDI)-induced AD model was established based on the previously described method with minor modification. 26 In brief, the back of the mice was shaved with an electric clipper and shaver. On the day after shaving, 50 μl of TDI (0.5% dissolved in acetone, Cat. No. 584-84-9, Tokyo Chemical Industry CO., LTD., Tokyo, Japan) was applied to the shaved dorsal skin and ears, respectively. Then, the same procedure was repeated twice a week for three weeks. Nontreatment (0 day) mice were collected immediately after hair removal.
The erythema/hemorrhage and scaling/dryness symptoms were scored twice a week as 0 (none), 1 (mild), 2 (moderate), and 3 (severe) based on the macroscopic criteria following the previously described method with minor modifications 24 (Tables 1 and 2). A total dermatitis score indicating clinical severity was defined as the sum of the individual scores (maximum score 6). Transepidermal water loss (TEWL) of Df-treated skin was measured in non-treatment, 4 and 7 days after Df-challenge using VAPO SCAN AS-VT100RS (Aschi Techno Lab, Tokyo, Japan) according to the manufacturer's guidelines. 27

| Histological analysis
Histological analysis was performed as previously described. 28 Briefly, skin tissues were excised and fixed with 4% buffered formaldehyde, embedded in paraffin, and sectioned at 5-μm thickness. The sectioned tissues were then stained by hematoxylin and eosin-stained (H&E) to measure the thickness of the epidermis. For the parameters, the average of 10 fields was estimated and used as the representative values. The number of mast cells was estimated by Toluidine blue staining. The number of mast cells was counted for 10 fields per mice (×200 magnification, single-blinded).
For immunohistochemistry, slides were subjected to antigen retrieval by Immunosaver (Nissin EM, Tokyo, Japan). Primary antibodies and Histofine Simple Stain MAX PO kit (for mouse antibodies; Cat. No. 414321, for rabbit antibody; Cat. No. 414341, Nichirei Biosciences, Tokyo, Japan) were used to visualize the protein expression in tissue. The following antibodies were used as pri- Immunostaining was performed following the manufacturer's instructions. Images were acquired using a BZ-X700 microscope (Keyence, Osaka, Japan), and the percentage of phospho-STAT3-and Ki67positive cells were calculated.

| Reverse transcription-quantitative polymerase chain reaction analysis
Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis was performed as previously described with minor modifications. 29 Table 3. Quantitative PCR conditions were as follows: 95°C for 10 min, followed by 40 cycles of 95°C for 10 s, 60°C for 15 s, and 72°C for 15 s. The fluorescence intensity was measured at the end of every extension cycle. Using β-2-microglobulin as internal control, relative mRNA quantities were obtained by E-method.

| ELISA
For cytokine evaluation, single-cell suspensions of LNs (5 × 10 5 cells/well) were incubated with mouse Dynabeads T-activator (Thermo Fisher Scientific, Inc.) for 24 or 96 h. IFNγ and TNFα concentrations in the supernatant were evaluated using ELISA according to the manufacturer protocols (R&D Systems). No. CP-690550, Pfizer Global Research and Development, Groton, CT) were pre-treated for 1 h before 10 ng/ml IL-6 treatment for 5 min. Since it is known that SM added to the extracellular can permeate the cell membrane and exert its function intracellularly, 6,31,32 MEFs were treated with or without 5 μM of SM (Cat. No. NS220103, Nagara Science Co., Ltd., Gifu, Japan) and BSM (crude fraction of sphingomyelin from butter, Genuine R&D, Fukuoka, Japan), and then with IL-6.

| SDS-PAGE and western blotting
Cells were collected and lysed in modified radioimmunoprecipitation assay (RIPA) buffer (

| Cell-cycle analysis
To analyze cell-cycle perturbation, DNA was stained with propidium iodide and the cells were analyzed by flow cytometry. 33 Cells were incubated with BSM and IL-6 for 48 h. We resuspended 1 × 10 6 cells in PBS and fixed them in ice-cold 70% ethanol for at least 6 h. Fixed cells were washed in PBS twice, and incubated with propidium iodide staining solution (50 µg/ ml propidium iodide, 5 µg/ ml RNase I, 137 mM NaCl, 2.7 mM KCl, 8.1 mM Na 2 PO 4 , and 1.47 mM KH 2 PO 4 ) at 37°C for 30 min. The DNA content of at least 10 000 cells/sample was analyzed using an EC800 Analyzer (Sony Biotechnology, Tokyo, Japan).

| Patients' data analysis
RNA-seq raw data were obtained by Gene Expression Omnibus under accession code GSE121212. Data sets were classified as AD patients (with lesions and without lesions) and healthy controls. Clinical data and RNA sequences are described elsewhere. 34 Reads Per Kilobase of exon per method was used to normalize the RNA-seq data set.

| Statistical analysis
The data are expressed as mean ± SD. The differences between the groups were analyzed by one or two-way analysis of variance, followed by Dunnett's multiple comparison test. Statistical significance was estimated at 5% and 1% levels of probability. The data were analyzed using Prism 9 (GraphPad Software, San Diego, CA, USA).

Df-induced dermatitis
To investigate the role of SM on the cutaneous barrier, Df-induced dermatitis was evaluated on the dorsal skin of WT, Sgms1, and Sgms2-KO mice. Df ointment barely induced dermatitis in WT mice; however, Sgms1-KO and Sgms2-KO mice showed a significant increase in cutaneous inflammation on 4-7 days post-Df application ( Figure 1A). The clinical dermatitis score analysis showed that WT mice were resistant to the Df-challenge, whereas the Sgms1-KO and Sgms2-KO mice were sensitive ( Figure 1B). In non-treatment WT mice, the immunohistochemical analysis showed that both SMS1 and SMS2 were highly expressed in the epidermis tissue. In contrast, we did not observe the expression of SMS1 or SMS2 on each KO mice ( Figure S1A). We observed that Df-treatment did not alter genes expression of Sgms1 and Sgms2 and the expression of Sgms1 was higher than that of Sgms2 in both time points ( Figure S1B) To assess the cutaneous barrier, we determined TEWL. Consistent with the symptoms of cutaneous inflammation, the value of TEWL was increased in Sgms1-KO and Sgms2-KO mice compared to WT mice at 4 and 7 days after Df-challenge ( Figure 1C). Moreover, the expression level of filaggrin, a major structural protein in the stratum corneum of the epidermis, 35 was also increased in Sgms1-KO and Sgms2-KO mice 7 days after Df-challenge ( Figure 1D). These results suggested that SM deficiency is associated with barrier dysfunction in the skin. Therefore, we performed the histological evaluation of cutaneous sections from WT, Sgms1-KO, and Sgms2-KO mice treated with Df ointment for 7 days. The epidermis of the skin lesions in Sgms1-KO and Sgms2-KO mice was significantly thickened compared to WT mice ( Figure 1E,F).
Next, we examined whether Sgms-deficient mice could upregulate skin inflammation. The gene expression of cytokines, which play key initiators of inflammation in epithelial tissues (IL-25, IL-33, and Thymic stromal lymphopoietin [TSLP]) or associated with cutaneous inflammation (IL-4 and IL-13) tended to increase in Sgms1-KO and Sgms2-KO mice, which was clearly in TSLP, IL-4 and IL-13 ( Figure S2A-E).
Although Sgms1/2-KO mice have shown severe cutaneous inflammation, we observe that immune cell infiltration was low in the Df-induced model. We then, investigate the effect of SMS1 and SMS2 on the skin immune response using the TDI-induced dermatitis model. Sgms1-KO mice showed severe inflammation compared to WT and Sgms2-KO mice ( Figure S3A). Also, Sgms1-KO mice showed a significant increase in inflammation score ( Figure S3B), TEWL ( Figure S3C), epidermal thickness ( Figures S3D,E), and immune cell infiltration ( Figure S3G) in comparison to WT and Sgms2-KO mice. Moreover, 21 days after TDI treatment, Sgms1-KO mice showed increased filaggrin expression compared to untreated mice, whereas the expression of it was decreased compared to WT and Sgms2-KO mice ( Figure S3F).
We examined whether Sgms-deficient mice could upregulate the immune reaction caused by TDI-induced allergic dermatitis. After TDI application, immune cells in LNs were analyzed by flow cytometry. The population of IgE-positive B cells and dendritic cells was increased in Sgms1-KO mice ( Figure S4A). The levels of IFNγ and TNFα were increased in Sgms1-KO mice after CD3/CD28 stimulation ( Figure S4B). Moreover, Sgms1-KO mice showed increased infiltration of mast cells in the dermis of dermatitis lesions ( Figure S4C).

STAT3 phosphorylation
Recent studies have reported the importance of the STAT3 pathway in developing AD and cutaneous inflammation. 17,36 Therefore, to investigate the role of SM in the STAT3 pathway, we evaluated the STAT3 phosphorylation levels in Sgms1-KO and Sgms2-KO mice after Df-ointment application. Seven days after Df treatment, the immunohistochemical analysis demonstrated an increased number of phosphorylated STAT3 (p-STAT3)-positive epidermal cells in Sgms1-KO and Sgms2-KO mice, which was more pronounced in Sgms1-KO mice (Figure 2A). Additionally, the quantitative results for p-STAT3-positive epidermal cells also revealed similar results in Sgms1-KO and Sgms2-KO mice ( Figure 2B). The increase in p-STAT3-positive cells was also observed in the allergic dermatitis model induced by TDI treatment, which was clearly in Sgms1-KO mice ( Figure S5). Therefore, we examined the functionality of p-STAT3 using one of the most cited markers for cell proliferation, Ki67. 37 The proliferation of basal keratinocytes in Sgms1-KO and Sgms2-KO mice skin after Df treatment was monitored by immunohistochemical analysis. Sgms1-KO and Sgms2-KO mice showed F I G U R E 1 Sgms knockout leads to progressive cutaneous inflammation in the Df-induced dermatitis model. Mice were administered Df ointment to induce dermatitis. Representative images of wild-type, Sgms1-KO, and Sgms2-KO mice at 0 (non-treatment), 4, and 7 days after Df treatment (A), and clinical dermatitis score. Scale bar = 1 cm (B). Analysis of transepidermal water loss of untreated and Df-treated mice after 7 days of the treatment (n = 3-4 mice/group) (C). Gene expression of filaggrin of untreated and Df-treated mice after 7 days of the treatment (n = 4-5 mice/group) (D). Histopathology was examined by H&E staining in untreated and Df-treated mice after 7 days of the treatment Scale bar = 100 µm (n = 3 mice/group) (E). Epidermal thickness was analyzed by measuring 10 points per mouse (n = 3 mice/ group) (F). Plots and bars are presented as the mean ± SD. *p < .05, **p < .01 increased Ki67-positive cells 7 days after Df application, which was prominent in Sgms1-KO mice ( Figure 2C,D). In addition, the gene expression levels of IL-6, IL-1β, IFNγ, and TNFα, which are known to the downstream of STAT3 in inflammation, [38][39][40] were also increased in Sgms1-KO and Sgms2-KO mice ( Figure S2F-I).

| Depletion of SM promotes the JAK/STAT3 pathway activated by IL6
To further investigate the molecular mechanism of STAT3 phosphorylation in Sgms-deficient mice, we generated immortalized MEF derived from WT and Sgms1-KO mice. Western blot analysis indicated that STAT3 was significantly phosphorylated in Sgms1-KO MEF at 0.01 and 0.1 ng/ml of IL-6, whereas STAT3 of WT MEF was barely phosphorylated under these low concentrations of IL-6 ( Figure  S6A). Then, we treated low (0.1 ng/ml) and high (10 ng/ml) concentration of IL-6 ( Figure 3A,B). While STAT3 was remarkably phosphorylated in Sgms1-KO in low IL-6 treatment, we did not observe a significant difference among WT and Sgms1-KO in a high concentration of IL-6.
To confirm the involvement of SMS1/2 in the STAT3 phosphorylation in keratinocytes, we employed the RNA interference of SGMS1 and SGMS2 in Human keratinocyte HaCaT cells. At 72 h after siRNA transfection, western blot analysis showed the downregulation of SMS1 and SMS2 expression ( Figure S7A,B). SGMS1 downregulation enhanced STAT3 phosphorylation in low concentrations (0.1 ng/ml) of IL-6, whereas SGMS2 downregulation was barely enhanced STAT3 phosphorylation under that of IL-6.
Janus kinase (JAK), an upstream protein to STAT3, has been widely implicated in inflammations involving dermatitis, particularly AD. 41 Therefore, we tested the hypothesis that inhibition of JAK would inhibit the phosphorylation of STAT3 using tofacitinib, a potent inhibitor of JAK1, JAK2, JAK3, and TYK2, which is currently under development as a therapeutic agent against inflammatory and autoimmune diseases. 42,43 To examine differences of JAK function between WT and Sgms1-KO MEF under conditions of equal STAT3 phosphorylation, we treated MEFs with 10 ng/ ml IL-6 after 0.1 µM tofacitinib pre-treatment. Tofacitinib attenuated the STAT3 phosphorylation in both WT and Sgms1-KO MEFs ( Figure 3C), suggesting the potential role of SMS1/2 in regulating the JAK/STAT3 pathway.

| SM controls the JAK/STAT3 pathway activated by IL-6
To confirm whether SM, the product of SMS1 and SMS2, can regulate JAK/STAT signaling, we examined the phosphorylation of STAT3 by treating SM and BSM pretreated cells with IL-6. Since 0.1 ng/ml IL-6 significantly increased STAT3 phosphorylation in Sgms1-KO MEF ( Figures S6A and 3A,B), we treated the MEF with 0.1 ng/ml IL-6 in the downstream experiments. WST assay showed IC 50 of SM and BSM were > 100 µM; we used 5 µM in this experiment. SM and BSM pretreatment suppressed STAT3 phosphorylation induced by IL-6 treatment in Sgms1-KO mice (Figures 3D and S6B,C). Consistent with this result, STAT3 phosphorylation was also decreased in the SM-treated HaCaT cell ( Figure S7C).
It has been reported that STAT3 is involved in cellular proliferation 44 ; therefore, as a next step, we examined the effects of SM on cell proliferation. It was observed that IL-6 treatment significantly increased cell proliferation consistent with STAT3 phosphorylation in Sgms1-KO MEFs, which was downregulated by SM and BSM treatment ( Figures 3E and S6D). However, BSM-treated MEF did not change the cell cycle distribution ( Figure S6E), which suggests that SM slows down cellular proliferation via STAT3 suppression.

| SM prevented cutaneous barrier dysfunction by Df-induced dermatitis in Sgms-KO mice
To examine the involvement of SM in dermatitis, gene expression of SGMS1/2 was analyzed in AD patients in silico. The expression levels of SGMS1 were significantly reduced in the skin of AD patients ( Figure 4A,B). Also, inflammatory cytokines (IL-5, IL-6, and IL-13) and chemokines (C-C motif chemokine [CCL]17 and CCL22) were increased in the skin of AD patients ( Figure 4C-G).
To further investigate whether SM is a therapeutic target for dermatitis, we applied BSM to the dorsal skin of Sgms1-KO and Sgms2-KO mice for one week before the Df-challenge. From Figures 3 and Figures S4B-D, there was no difference between SM and BSM in the effect on STAT3 regulation. Therefore, BSM, water-soluble and less damaging to the skin, was treated to these mice. As shown in Figure 5A, BSM suppressed cutaneous inflammation in Sgms1-KO and Sgms2-KO mice. Clinical dermatitis score was also reduced in BSM-treated mice compared to the control mice ( Figure 5B). Moreover, the BSM treatment prevented the increase in TEWL in cutaneous inflammation ( Figure 5C) as well as the increase in epidermal thickness ( Figure 5D,E). Finally, dermatitis was suppressed and the number of phospho-STAT3-expressing cells in epidermal tissues was reduced after BSM treatment ( Figure  5F,G). These results indicated that SM plays an important role in maintaining the cutaneous barrier and preventing dermatitis via STAT3 suppression.

F I G U R E 3
Sgms knockout enhanced STAT3 phosphorylation by IL-6 in mouse embryonic fibroblast (MEF). MEF derived from wildtype, and Sgms1-KO mice were treated with IL-6 to induce inflammation. Expression of p-STAT3 treated with 0.1 ng/ml of IL-6 for the indicated times (n = 3/group) (A). Expression of p-STAT3 treated with 10 ng/ml of IL-6 for the indicated times (n = 3/group) (B). Expression of p-STAT3 treated with 0.1 µM tofacitinib for 1 h before treatment with 10 ng/ml IL-6 for 5 min (n = 3/group) (C). Expression of p-STAT3 treated with 5 µM SM for 12 h before 0.1 ng/ml IL-6 treatment for 12 h (n = 4-6/group) (D). Growth rates in 0.1 ng/ml IL-6 and 5 µM SM treatment for 72 h (n = 3/group) (E). The relative band intensity of phosphorylated proteins was normalized to that of total proteins. Plots and bars are presented as the mean ± SD. *p < .05, **p < .01

| DISCUSSION
Skin, the largest organ of the body, contains a strong barrier preventing the penetration of external aggressors, such as bacteria, fungi, and chemicals, 45 and the dysfunction of the cutaneous barrier has been implicated in chronic inflammatory skin diseases, such as AD. In the present study, we demonstrated that SM, a sphingolipid expressed in the plasma membrane, plays a key role in maintaining the cutaneous barrier. It was shown that SM treatment ameliorated the STAT3 hyper-phosphorylation in Sgms-deficient mice, indicating the potential of SM in preventing the cutaneous barrier dysfunction and reducing the aggravation of dermatitis.
The stratum corneum consists of corneocytes surrounded by multilamellar lipid membranes that prevent excessive water loss from the body and entrance of undesired substances and plays a pivotal role in forming a proper physiological cutaneous barrier. Several investigators have reported that the skin of patients with AD has a decreased proportion of ceramide, a major intercellular lipid in the stratum corneum. 46,47 SM, produced from ceramide via SMS, is an essential lipid enriched in the plasma membranes of animal cells, where it regulates membrane properties and many intracellular signaling processes. 48,49 Sgms1 and Sgms2 double KO mice exhibited viviparous lethality (data not shown); therefore, in this study, Sgms1-KO, and Sgms2-KO mice were employed.
Our study found exacerbation of dermatitis in Sgmsdeficient mice, which was more pronounced in Sgms1-KO mice than Sgms2-KO ( Figures 1E,F and S3G). The difference in the dermatitis response between Sgms1-KO and Sgms2-KO could be due to the difference in SM content level in tissue type. SMS1 expresses almost all the tissues, whereas SMS2 varies expression depending on the tissue, and is known to be abundant in liver, small intestine, colon, and kidney but not in the immune cell. 50 As shown in Figures S3 and S4, dermatitis was exacerbated only in Sgms1-KO mice with extensive infiltration of immune cells. Several studies have reported that SMS1 produces 80% of the cellular SM and SMS2 produces 20% and reduction of SM content in Sgms1-or Sgms2-deficient mice. In addition, the total tissue SM in Sgms1-KO mice was reported to be significantly lower than that of Sgms2-KO mice. 10,11,51,52 Moreover, induction of dermatitis hardly altered the gene expressions of Sgms1 and Sgms2 ( Figure  S1B). These results suggest that both Sgms1/2 are constitutively expressed in skin tissues and that SM levels are involved in the maintenance of skin barrier function.
Aberrant lipid organization in the lipids increases TEWL in AD skin. 47,53 Consistent with these reports, our data showed that TEWL levels were increased in Sgmsdeficient mice together with cutaneous inflammation ( Figures 1C and S3C). It has been reported that changes in the components of the lipid membrane in the epidermis disrupt the cutaneous barrier, 54 which strongly suggests that the components of the lipid membrane affect the cutaneous barrier and subsequent pathology of dermatitis. Additionally, dietary supplementation with a concentrate of milk phospholipids, including SM, has been shown to significantly increase the hydration levels of the skin. 55 Furthermore, SMS has been shown to be associated with inflammation, including pneumonia and colitis. 11,12 Consistent with these reports, in our experiments, the Sgms deficient mice were unable to maintain their membrane properties leading to the onset of cutaneous barrier dysfunction and dermatitis development. IL-25, IL-33, and TSLP are the cytokines released from the epidermis when the cutaneous barrier is damaged and activates immune cells such as mast cells, dendritic cells, and Th2 cells. [56][57][58][59][60] The levels of mast cell and Th2 cytokines, IL-4 and IL-13, were also increased along with inflammation ( Figures S2D,E and  S4B). In addition, Sgms1-KO mice sometimes showed inflammatory responses even before the induction of dermatitis. As shown in Figure S2A-C, the mRNA levels of IL-25, IL-33, and TSLP tend to increase in Sgmsdeficient mice both in non-treated and Df-treated mice, which was clearly in TSLP. These data suggested that the slightest stimulation (such as depilation by hair remover) initiates dermatitis in SM-deficient skin. Further investigation with cutaneous-specific conditional Sgms1 and Sgms2 double KO mice is necessary.
Filaggrin is a key component of the stratum corneum for maintaining a normal cutaneous barrier. Decreased expression of filaggrin in the skin and loss of function mutants of the filaggrin gene have been described in AD, 61,62 whereas the increased filaggrin expression in human and murine keratinocytes attenuated the development of ADlike cutaneous inflammation. 63 The filaggrin expression of spontaneous AD mouse models such as NC/Nga should be suppressed as the symptom progressed. Mechanism of dermatitis development and the recovery process of C57BL/6N mice and the spontaneous AD model might be F I G U R E 5 BSM prevented cutaneous barrier dysfunction in the Df-induced dermatitis model. Mice were treated with 10% BSM solution for 7 days before Df ointment. Representative images of Sgms1-KO and Sgms2-KO mice in non-treatment, Df treatment for 7 days, and BSM treatment for 7 days before Df ointment (A), and clinical dermatitis score (Sgms1-KO mice n = 3-4/group and Sgms2-KO mice n = 4-7/group) (B). Scale bar = 1 cm. Analysis of transepidermal water loss in non-treated (Sgms1-KO mice n = 5 and Sgms2-KO mice n = 3), Df-treated (for 7 days) (Sgms1-KO mice n = 3 and Sgms2-KO mice n = 5), and BSM-treated (for 7 days before Df ointment) (Sgms1-KO mice n = 3 and Sgms2-KO mice n = 4) mice (C). Histopathological examination by H&E staining in non-treated, Df-treated (for 7 days), and BSM-treated (for 7 days before Df ointment) mice. Scale bar = 100 µm (D). Epidermal thickness was analyzed by measuring 10 points per mouse (n = 3 mice/group) (E). Histopathology was examined by immunohistochemistry with p-STAT3 in non-treated, Df-treated (for 7 days), and BSM-treated (for 7 days before Df ointment) mice. Scale bar = 100 µm (F). Phospho-STAT3 (p-STAT3)-positive epidermal cells were analyzed by counting the number of p-STAT3-positive nuclei in 100 nuclei per mouse; the percentage was calculated (n = 3 mice/group) (G). Plots and bars are presented as the mean ± SD. *p < .05, **p < .01 different, therefore, further analysis has to be performed in the future study.
The activation of STAT3 signaling occurs in the skin of AD patients, 64,65 wherein phosphorylated STAT3 in keratinocytes has been shown to play an important role in the onset of AD by promoting the expression of inflammatory cytokines. In this study, hyper-phosphorylation of STAT3 in the epidermis was observed in Sgms-deficient mice with dermatitis (Figures 2A,B and S5); in addition, STAT3downstream genes were also increased ( Figure S2F-I).
In addition to the mice dermatitis model, we performed in vitro experiments using Sgms1-KO MEF, which had a more robust dermatitis phenotype than Sgms2-KO mice and lower expression in patient data. Consistent with our findings in mouse experiments, the phosphorylated STAT3 was also increased in MEFs derived from Sgms-deficient mice treated with a low concentration of IL-6 ( Figure 3A), indicating that the cellular sensitivity to IL-6 was increased by Sgms deficiency. These findings are consistent with the findings of a recent study demonstrating that the phosphorylated STAT3 levels reflect inflammation severity. 66 The experiment with MEF indicated that Sgms deficiency might participate in STAT3 phosphorylation via effects on JAK activation. JAK has been reported as a signal transducer for STAT3 in various inflammatory processes, including dermatitis. In this study, tofacitinib, a clinical JAK inhibitor, inhibited STAT3 phosphorylation induced by IL-6, indicating that JAK functions upstream of STAT3 in the present condition. It has been reported that tofacitinib directly enhances the expression of structural proteins in the epidermis, such as filaggrin, and is expected to effectively reduce inflammation and improve the cutaneous barrier via the JAK/STAT pathway in AD. 65 Therefore, our findings are that the JAK/STAT pathway is involved in dermatitis in Sgms-deficient mice. Furthermore, the BSM, used as cosmetic material, suppressed STAT3 phosphorylation in vitro and in vivo ( Figures 3D,E, 5, S6C-E and S7C) and prevented dermatitis in Sgms-deficient mice ( Figure 5). These results suggest SM is involved in the maintenance of skin barrier function through the inhibition of STAT3 and the clinical applicability of SM to treat dermatitis.
In summary, the present study demonstrates that the cutaneous barrier dysfunction in Sgms-KO mice sensitizes the inflammation to epicutaneous antigen. In addition to the skin inflammation triggered by the entrance of undesired substances, excessive phosphorylation of STAT3 in the cytoplasm causes the development of dermatitis. Our results show that SM is essential for maintaining the cutaneous barrier and is regulated by the JAK/STAT pathway. Overall, these findings strongly suggest that SM could be an important therapeutic target in treating dermatitis ( Figure 6).

ACKNOWLEDGMENTS
We would like to thank Editage (www.edita ge.jp) for English language editing. This work was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Numbers 17K08140 [TY] and 19K20452 [MN].

AUTHOR CONTRIBUTIONS
All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Mariko Komuro, Masaki Nagane and Tomoki Fukuyama. The first draft of the manuscript was written by Mariko Komuro and Masaki Nagane and Tomoki Fukuyama commented on the previous versions of the manuscript. All authors read and approved the final manuscript. Conceptualization: Tadashi Yamashita. Methodology: Mariko Komuro, Masaki Nagane, Tomoki F I G U R E 6 Schematic representation of the cutaneous barrier dysfunction via activation of the STAT3 pathway in the absence of sphingomyelin suppressing signal transducer and activator of transcription 3 signaling. J Allergy Clin Immunol.

SUPPORTING INFORMATION
Additional supporting information may be found in the online version of the article at the publisher's website.