Before the experiments, female BALB/c mice (6-7 weeks old) were purchased from Japan SLC (Hamamatsu, Japan) and maintained in a specific-pathogen-free animal facility at NIBIOHN (National Institutes of Biomedical Innovation, Health and Nutrition, Ibaraki, Osaka, Japan). Mice were euthanized by cervical dislocation under anesthesia with isoflurane (AbbVie, North Chicago, Illinois, United States). All experiments were conducted in accordance with the guidelines of the Animal Care and Use Committee and the Committee on the Ethics of Animal Experiments at NIBIOHN.

Retinol-induced ICD was induced as described previously, with some modifications (Lee et al., 2010; Kim et al., 2012). Briefly, both ears of each mouse were treated topically once daily for 4 consecutive days with 10 µL of 0.05% (w/v) all-trans-retinol (Sigma-Aldrich, St. Louis, Missouri, United States) dissolved in 4:3:3 dimethyl acetamide (Sigma-Aldrich), acetone (Nacalai Tesque, Kyoto, Japan), and ethanol (Nacalai Tesque). Mead acid (10 µg/10 μL, Cayman Chemical, Ann Arbor, Michigan, United States), oleic acid (10 µg/10 μL, Cayman Chemical), or vehicle control 3:1 ethanol in phosphate-buffered saline (PBS) was applied topically to both sides of the ears 45 min after each of the inflammatory insults. Ear thickness was evaluated with a micrometer MDC-25MJ 293–230 (Mitsutoyo, Kawasaki, Japan).

In experiments using receptor antagonists, GSK0660 (100 μg/10 μL, Sigma-Aldrich), GW6471 (100 μg/10 μL, Sigma-Aldrich), or vehicle control 1:1 dimethyl sulfoxide (DMSO, Nacalai Tesque) and 50% (v/v) ethanol in PBS was applied topically to both sides of the ears. Thirty minutes after treatment with the antagonist, mead acid or vehicle control was applied topically, and 45 min later retinol was applied topically at the dose rate described above. These treatment sequences were performed once daily for 4 consecutive days. In experiments using kinase inhibitors, SB202190 (Sigma-Aldrich), or U1026 (Sigma-Aldrich), or SP600125 (Sigma-Aldrich), or vehicle control 1:1 DMSO and 50% (v/v) ethanol in PBS were applied at 20 mM/10 μL topically on both sides of the ears 45 min after topical application of the retinol at the dose rate described above. These treatment sequences were continued once daily for 4 consecutive days. Ear swelling (Δµm) was calculated as (Ear thickness [µm] on the indicated day)—(Ear thickness [µm] before the first retinol application).

Cell isolation and flow cytometry were performed as described previously (Nagatake et al., 2018). In brief, murine ear samples were digested, with stirring, for 60 or 90 min at 37°C with 2 mg/mL collagenase (Wako Pure Chemicals, Osaka, Japan) in RPMI 1640 medium (Sigma-Aldrich) containing 2% (v/v) newborn calf serum (Equitech-Bio, Kerrville, Texas, United States). We treated the samples with collagenase for 90 min to analyze immune cells and for 60 min to sort keratinocytes. Cell suspensions were filtered through a 70-µm cell strainer (BD Biosciences, Franklin Lakes, New Jersey, United States) and stained with FITC (fluorescein isothiocyanate)–anti-mouse-Ly6G antibody (BioLegend, San Diego, California, United States; 1:100), APC (allophycocyanin)–Cy7 anti-CD11b antibody (BioLegend; 1:100), PE (phycoerythrin)–anti-CD31 antibody (BD Biosciences; 1:100), FITC–anti-CD34 antibody (BD Biosciences; 1:100), APC–anti-CD49f antibody (BioLegend; 1:100), PE–anti-Ki-67 antibody (BioLegend; 1:100), or BV (Brilliant Violet)-421–anti-CD45 antibody (BioLegend; 1:100). Murine keratinocytes were isolated as the CD45^– CD31^– CD34^– CD49f⁺ cell fraction, as described previously (Saika et al., 2021). Samples were analyzed by using a MACSQuant (Miltenyi Biotec, Bergisch Gladbach, Germany) or FACSAria (BD Biosciences) cell sorter, and cells were isolated with a FACSAria. Data were analyzed by using Flowjo 9.9 (Tree Star, Ashland, Oregon, United States).

Histological analysis was performed as described previously (Nagatake et al., 2018). The mice were sacrificed on day 4 of the experiment, after which ear samples were collected immediately. Frozen tissue sections (6 μm thick) were prepared by cryostat (Leica, Wetzlar, Germany, CM3050S) and used for hematoxylin and eosin staining and immunohistological analysis. The following antibodies were used in the immunohistological analysis: anti-cytokeratin 10 (K10) antibody (Abcam, Cambridge, United Kingdom), FITC–anti-Ly6G antibody, PE–anti-Ki-67 antibody, and APC–anti-CD49f antibody. AF (Alexa-Fluor)-488–anti-rabbit IgG (Thermo Fisher Scientific, Waltham, Massachusetts, United States) was used as a secondary antibody to detect K10. Cell nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI, AAT Bioquest, Sunnyvale, California, United States). Tissue sections were examined under a fluorescence microscope (model BZ-9000, Keyence, Osaka, Japan). The staining area of images was analyzed by using the Analyzer software for the BZ-X700 (Keyence).

The Indigo Biosciences (State College, Pennsylvania, United States) Reporter Assay System was used for the luciferase reporter assay. Fatty acids were tested for their ability to activate nuclear receptors by using human PPARα and PPARβ reporter assay systems (Indigo Biosciences) followed by the procedure described in the kits. In brief, reporter cells expressing a hybrid receptor composed of the Gal4 DNA-binding domain fused to the ligand-binding domain of the specific nuclear receptor, together with the firefly luciferase reporter gene, were provided with the reporter assay systems. Several concentrations (0.3, 3, and 30 µM) of mead acid and oleic acid were used for the assay. Oleic acid was used as a control for mead acid. After 24 h of incubation (37°C in 5% CO₂) of these mixtures with the tested compounds, the activities of the nuclear receptors were quantified as relative light units by measuring the emission of light with a microplate luminometer (Arvo X2, Perkin Elmer, Waltham, Massachusetts, United States).

HaCaT cells (immortalized human keratinocytes) were obtained from CLS Cell Lines Service (Eppelheim, Germany) (Boukamp et al., 1988) and cultured as described previously (Saika et al., 2021). Cultured HaCaT cells were seeded at a concentration of 5 × 10⁵ cells/10 mL in Dulbecco’s modified Eagle’s medium high glucose (Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS, Gibco) and 100 U/mL penicillin with 100 μg/mL streptomycin (Nacalai Tesque) and incubated for 24 h at 37°C in 5% CO₂. The medium was then replaced with Dulbecco’s modified Eagle’s medium without FBS, and the cells were treated with 3 µM mead acid or vehicle (ethanol) for 30 min before stimulation with HB-EGF (1 ng/mL, PeproTech, Cranbury, New Jersey, United States) for 1 h at 37°C in 5% CO₂.

HaCaT cells were harvested by being scraped in 300 µL of RIPA (Radioimmunoprecipitation Assay) lysis buffer (EMD Millipore Corporation, Burlington, Massachusetts, United States) containing protease cocktail inhibitor (Sigma-Aldrich) and PhosSTOP phosphatase inhibitor (Roche, Basel, Switzerland). The protein in the samples was quantified by using a BCA (bicinchoninic acid) protein assay kit (Thermo Fisher Scientific) in accordance with the instructions in the kit. An equal amount of protein was separated on NuPAGE 4%–12% Bis-Tris gels (Thermo Fisher Scientific), and transferred to polyvinylidene difluoride membranes (EMD Millipore Corporation). The antibodies used were anti-phospho-p38 MAPK and anti-p38 MAPK polyclonal antibodies (Cell Signaling, Danvers, Massachusetts, United States) and horseradish-peroxidase-conjugated donkey anti-rabbit IgG (BioLegend). Western blot signals were analyzed by using an image analyzer (LAS-4000, ImageQuant, GE Healthcare, Tokyo, Japan).

After daily application of retinol or vehicle for 4 days, ear samples were harvested in XXTuff microvials (BioSpec Products, Bartlesville, Oklahoma, United States), disrupted by using stainless-steel beads (φ 48 × 1, φ 32 × 4, TOMY, Tokyo, Japan) at 4,800 rpm for 10 s, and then pulsed five times by using a tissue homogenizer (Precellys 24, Bertin Instruments, Montigny-le-Bretonneux, France). Total RNA was isolated by using the ReliaPrep RNA Tissue Miniprep System (Promega, Madison, Wisconsin, United States). In some experiments, keratinocytes were isolated from ear samples; total RNA was prepared by using Sepazol reagent (Nacalai Tesque) and chloroform (Nacalai Tesque), precipitated with 2-propanol (Nacalai Tesque), washed with 75% (v/v) ethanol (Nacalai Tesque), and treated with DNase Ι (Promega).

cDNA was prepared from samples by using a cDNA Synthesis Kit (Invitrogen, Carlsbad, California, United States). Quantitative RT-PCR analysis was performed with a LightCycler 480 II PCR platform (Roche) with FastStart Essential DNA Probes Master reaction mix for PCR (Roche) or SYBR Green I Master reagents (Roche).

Gene expression levels were normalized to that of the β-actin gene Actb. The following primer sequences were used for PCR with FastStart Essential DNA Probes Master for mouse: Cxcl1 forward, 5′-gac​tcc​agc​cac​act​cca​ac-3´; Cxcl1 reverse, 5′-tga​cag​cgc​agc​tca​ttg-3′; Cxcl2 forward, 5′-aaa​atc​atc​caa​aag​ata​ctg​aac​aa-3′; Cxcl2 reverse, 5′-ctt​tgg​ttc​ttc​cgt​tga​gg-3′; Hbegf forward, 5′-tct​tct​tgt​cat​cgt​ggg​act-3′; Hbegf reverse, cac​gcc​caa​ctt​cac​ttt​ct-3′; Actb forward, 5′-aag​gcc​aac​cgt​gaa​aag​at-3′; and Actb reverse, 5′-gtg​gta​cga​cca​gag​gca​tac-3′. The following primer sequences were used for PCR with SYBR Green I Master reagents for human: DUSP1 forward, 5′-agt​acc​cca​ctc​tac​gat​cag​g-3′; DUSP1 reverse, 5′-gaa​gcg​tga​tac​gca​ctg​c-3′; DUSP6 forward, 5′-gaa​atg​gcg​atc​agc​aag​acg-3′; DUSP6 reverse, 5′-cga​cga​ctc​gta​tag​ctc​ctg-3′; MKK3 forward, 5′-gac​tcc​cgg​acc​ttc​atc​ac-3′; MKK3 reverse, 5′-ggc​cca​gtt​ctg​aga​tgg​t-3′; MKK4 forward, 5′-tgc​agg​gta​aac​gca​aag​ca-3′; MKK4 reverse, 5′-ctc​ctg​tag​gat​tgg​gat​tca​ga-3′; ACTB forward, 5′- cat​gta​cgt​tgc​tat​cca​ggc-3′; and ACTB reverse, 5′-ctc​ctt​aat​gtc​acg​cac​gat -3´.

Cell viability was calculated by using a CellTiter 96 Non-Radioactive Cell Proliferation Assay kit (Promega) in accordance with the manufacturer’s instructions. Briefly, HaCaT cells were seeded at 1 × 10⁴ cells/well on 96-well plates and then incubated for 24 h at 37°C in 5% CO₂. Mead acid or oleic acid was added, and the plates containing fatty acids in each concentration of 300 pM to 300 μM were incubated for 1 h or 24 h. After each exposure time, the culture medium was removed, and 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide dye solution was added to each well. The plates were then incubated for 4 h at 37°C in 5% CO₂, and solubilization/stop solution was added. After an overnight incubation, the absorbance at OD₅₇₀ was measured by using an iMark microplate reader (Bio-Rad, Hercules, California, United States). Ethanol at 0.2% (v/v) in the culture medium was used as a vehicle control. The percentage of viable cells was calculated by using the following formula: (%) = [100% × (sample OD₅₇₀)/(vehicle control OD₅₇₀)].

Data were analyzed by using the non-parametric Kruskal–Wallis test followed by Dunn’s multiple comparison test, the Mann-Whitney U-test, or Welch’s t-test (Prism 6, GraphPad Software, San Diego, California, United States). A p-value of less than 0.05 was considered significant.

Given that topical retinol treatment causes epidermal thickening (Lee et al., 2010; Kong et al., 2016), we first evaluated the inflammatory signs in a murine model of retinol-induced ICD. In this model, retinol was applied topically to both sides of the ear for 4 consecutive days and ear thickness (and hence swelling) was measured on days 1, 2, 3, and 4. Ear swelling increased in a time-dependent manner in the retinol-treated mice but not in the untreated naïve mice (Figure 1A). In addition, on days 2–4, mead acid treatment reduced the ear swelling induced by retinol treatment compared with vehicle treatment (Figure 1A). Histological analysis of ear sections of vehicle-treated mice showed increases in ear thickness, the thickness of the epidermal layer, and cell infiltration into the dermis compared with those in naive mice (Figure 1B). However, these inflammatory signs were ameliorated in ears cotreated with mead acid (Figure 1B). Because mead acid is a metabolite derived from oleic acid, we also examined whether oleic acid also had anti-inflammatory activity, but no anti-inflammatory activity was detected (Figure 1C), indicating that the suppressive effect in retinol-induced ICD was a particular characteristic of the mead acid treatment.

We previously reported that intraperitoneal injection of mead acid inhibited neutrophil chemotaxis into the site of inflammation in a DNFB-induced allergic contact hypersensitivity murine model (Tiwari et al., 2019). In addition, we showed that mead acid acted directly on neutrophils and inhibited neutrophil migration by suppressing their pseudopod formation and LTB₄ production. Neutrophil accumulation in the dermis has been reported in both retinol dermatitis and DNFB-induced allergic contact hypersensitivity (Varani et al., 2008). In light of our previous findings, we examined whether mead acid regulated neutrophils in retinol-induced ICD. As we showed in our previous report (Tiwari et al., 2019), mead acid reduced the increase in neutrophil numbers in inflamed ear skin (Figures 2A, B). Immunohistological analysis revealed that the abundance of Ly6G⁺ neutrophils in the dermis was lower in the ear sections of mice that had received mead acid co-treatment (Figure 2C). These results indicated that mead acid inhibited neutrophil infiltration into the inflamed skin in retinol-induced ICD.

Neutrophil recruitment is promoted by the indirect mechanisms of neutrophil chemoattractants such as Cxcl1 and Cxcl2 (Saika et al., 2021). The major source of neutrophil-attracting chemokines is keratinocytes (Lee et al., 2013), and neutrophil numbers after induction of inflammation are reduced by inhibition of the expression of neutrophil-attracting chemokines on keratinocytes (Cattani et al., 2006; Saika et al., 2021). Therefore, we analyzed the expression levels of Cxcl1 and Cxcl2 on keratinocytes fractionated from murine ears to reveal the indirect neutrophil-regulating activity of mead acid. We found that the increase in expression of Cxcl1 and Cxcl2 on keratinocytes sorted from murine ear skin after retinol stimulation was inhibited by mead acid co-treatment (Figure 2D). Thus mead acid inhibited neutrophil infiltration into the inflamed skin by both direct mechanisms, as shown previously (Tiwari et al., 2019), and indirect mechanisms.

In the retinol-induced ICD murine model, epidermal hyperplasia and the consequent increase in local skin thickness are accompanied by the enhancement of keratinocyte proliferation (Chapellier et al., 2002). Because we found that mead acid regulated not only neutrophil function but also keratinocyte function, we used histological analysis to evaluate the degree of epidermal hyperplasia. To this end, we stained ear sections for K10, which is expressed on the stratum spinosum. Whereas retinol treatment increased the thickness of the K10-positive keratinocyte layer, co-treatment with mead acid reduced the thickness of this layer (Figures 3A, B). We next evaluated the expression of Ki-67, a marker of proliferation, in keratinocytes. When we segregated basal keratinocytes by CD49f expression as a marker of epidermal stem cells, Ki-67-expressing cells were predominantly located in the upper part of the basal keratinocyte layer in naïve mice. Retinol treatment increased the numbers of these Ki-67-expressing cells, but mead acid co-treatment decreased the Ki-67-positive cell count (Figure 3C). Flow cytometric analysis confirmed the results by showing that the abundance of Ki-67-positive cells among CD45^– CD31^– CD34^– CD49f⁺ keratinocytes was increased by retinol treatment, whereas the increase was suppressed by mead acid co-treatment (Figure 3D). These results indicated that mead acid inhibited retinol-induced epidermal hyperplasia by inhibiting keratinocyte proliferation.

To identify the functional receptor of mead acid in its anti-inflammatory activities in retinol-induced ICD, we next performed a reporter assay. PPARs are receptors for long-chain free fatty acids and are associated with the reduction of inflammation in conditions such as dermatitis, colitis, inflammatory bowel disease, diabetes, and asthma (Sertznig et al., 2008; Staumont-Sallé et al., 2008; Echeverria et al., 2016). Therefore, we evaluated PPAR activity. We found that, unlike oleic acid, mead acid activated PPARα and PPARβ/δ in a concentration-dependent manner (Figure 4A). We confirmed that mead acid and oleic acid were not toxic to HaCaT cells at any concentration from 300 pM to 300 μM, including the 0.3–30 μM concentrations used in the in vitro assay. (Supplementary Figures S1A, B). As mead acid activated PPARγ to almost the same level as those of other free long-chain fatty acids (data not shown), we aimed to evaluate only PPARα and PPARβ/δ to determine whether they acted as functional receptors of mead acid for the amelioration of retinol-induced ICD. To this end, we evaluated the effects of mead acid by topical application of the respective receptor antagonists before mead acid treatment. We found that the inhibitory effect of mead acid on epidermal hyperplasia was cancelled by co-treatment with GW6471 (a PPARα antagonist) but not with GSK0660 (a PPARβ/δ antagonist) (Figures 4B, C). Moreover, the modulating effect of mead acid on the increase in abundance of Ki-67-positive cells among keratinocytes was cancelled by co-treatment with GW6471 but not by treatment with GSK0660 (Figure 4D). In addition, the ability of mead acid to prevent neutrophil infiltration via the control of Cxcl1 and Cxcl2 expression on keratinocytes and of neutrophil numbers in inflamed skin was cancelled by co-treatment with GW6471 (Figures 4E, F). Consequently, co-treatment with GW6471 led to a lack of inhibition of ear swelling by mead acid (Figure 4G). These results indicated that mead acid acted as a ligand of both receptors, PPARα and PPARβ/δ, but that PPARα was a functional receptor for its inhibition of skin inflammation in retinol-induced ICD.

Retinol is known to enhance MAPK pathways, which play essential roles in the regulation of cell proliferation and in the production of inflammatory cytokines and chemokines (Yen et al., 1998; Lee et al., 1999; Yu et al., 2003; Lee et al., 2008; Piskunov and Rochette-Egly, 2012). The underlying signaling that accelerates retinoid-induced inflammation is not fully understood. Therefore, by using specific inhibitors, we first evaluated which major signaling pathway, namely p38 MAPK, MAPK/extracellular signal-regulated kinase (ERK), or c-Jun amino-terminal kinase (JNK), was involved in retinol-induced keratinocyte abnormalities.

To reveal the pathways associated with the anti-inflammatory properties of mead acid, we evaluated epidermal hyperplasia and keratinocyte proliferation in response to treatment with kinase inhibitors after retinol treatment. We found that the K10-staining layer of the epidermis and the area stained with K10 were decreased by treatment with SB202190 (a p38 MAPK inhibitor) and U0126 (a MAPK/ERK inhibitor), but not by treatment with SP600125 (a JNK inhibitor) (Figures 5A, B). In addition, the percentage increase in Ki-67-positive keratinocytes and the increases in expression of Cxcl1 and Cxcl2 on keratinocytes were reduced only by SB202190 treatment (Figures 5C, D). Consistent with this finding, SB202190 treatment after retinol application reduced the number of neutrophils at the site of inflammation (Figure 5E). Moreover, although U0126 treatment had no effects on keratinocytes (i.e., cell proliferation and chemokine production) (Figures 5C, D), its application decreased the numbers of neutrophils at the site of inflammation (Figure 5E). Reflecting these suppressive effects on keratinocyte and neutrophil activities, ear swelling was reduced by treatment with either SB202190 or U0126, but not by SP600125 treatment (Figure 5F).

Secretory granules from neutrophils contain proteases that induce further injury to, and activation of, keratinocytes. Our finding that the topical application of U0126 reduced neutrophil infiltration (Figure 5E) suggested that the MAPK/ERK pathway regulates neutrophil function and thus reduces keratinocyte hyperproliferation. In contrast, given that topical application of SB202190 reduced keratinocyte activities such as proliferation and chemokine production, we considered that the p38 MAPK pathway is essential for induction of the keratinocyte abnormalities induced by retinol. In addition, we confirmed that further inhibition of keratinocyte abnormalities such as enhancement of keratinocyte proliferation and production of neutrophil chemoattractants, and ear swelling was not observed upon co-treatment with SB202190 and mead acid (Supplementary Figures S2A–C). These results suggested that the p38 MAPK pathway is a candidate for the target pathway regulated by mead acid.

By using the human keratinocyte cell line HaCaT, we next examined whether mead acid inhibited the p38 MAPK pathway. HB-EGF is an inducer of keratinocyte hyperplasia, and Hbegf expression on keratinocytes is enhanced through RAR and RXR heterodimer activation, which is activated by retinol (Rittie et al., 2006). We found that mead acid did not reduce Hbegf expression on keratinocytes after retinol treatment (Supplementary Figure S3), suggesting that the inhibitory effect of mead acid was downstream of HB-EGF signaling. Therefore, we next analyzed the phosphorylation of p38 by stimulating HaCaT cells with HB-EGF for 30 min after mead acid treatment. Western blot analysis showed that the expression level of p38 was not changed by mead acid treatment, whereas the expression of phospho-p38 was significantly reduced by mead acid treatment (Figures 6A, B). To identify the mechanism of reduction of phospho-p38, we focused on two enzymes that catalyze the phosphorylation of p38 and are known to activate p38 MAPK; and protein phosphatases that catalyze dephosphorylation and are negative regulators of phospho-p38 (Kidger and Keyse, 2016). We confirmed that the expression levels of the MAPK kinases MKK3 and MKK4 were not affected by mead acid treatment after retinol application (Supplementary Figure S4). The largest group of protein phosphatases that specifically regulates MAPK activity in mammalian cells is the dual-specificity phosphatase (DUSP) family of phosphatases, which are known as MAPK phosphatases (MKPs) (Kidger and Keyse, 2016; Chen et al., 2019). We next evaluated MKP gene expression on HaCaT cells after stimulation with HB-EGF. Mead acid upregulated the expression of MKP1/DUSP1 and MKP3/DUSP6 compared with vehicle treatment (Figure 6C). These results suggested that mead acid resolves p38 phosphorylation by enhancing MKP expression in keratinocytes, consequently reducing p38 MAPK activation.