For lipidomics analysis, we purchased male germ-free (GF) mice and control ICR mice (age, 6 weeks) from Japan SLC (Hamamatsu, Japan); these mice were maintained for 2 months on chemically defined diets containing 4% (wt/wt) dietary oil comprising soybean oil or linseed oil (Oriental Yeast, Tokyo, Japan). GF mice were housed under GF conditions at Oriental Bioservice (Kyoto, Japan), and control mice were housed under specific-pathogen-free (SPF) conditions at the NIBIOHN (Osaka, Japan).

For the contact hypersensitivity model, we obtained female wild-type C57BL/6 mice (age, 6–8 weeks) from SLC (Shizuoka, Japan) and housed them in an SPF animal facility at NIBIOHN for at least 1 week before their use in experiments. In this study, female mice were chosen for the contact hypersensitivity model due to their lower aggression levels compared to males (Schwarz et al., 2023). These mice had ad libitum access to distilled water and a commercially available standard diet (FR2, Funabashi Farm, Chiba, Japan) under conditions of 22–24°C, 50%–60% humidity, and a 16:8-h light:dark cycle. Mice were euthanized by cervical dislocation under anesthesia with isoflurane (AbbVie Inc., North Chicago, Illinois, USA). All experiments were performed in accordance with the guidelines of the Animal Care and Use Committee and the Committee on the Ethics of Animal Experiments at NIBIOHN.

The model was generated as described previously (Nagatake et al., 2018; Saika et al., 2021). Briefly, on day 0 the shaved abdominal skin of C57BL/6 mice was treated with 25 μL of 0.5% (vol/vol) DNFB (Nacalai Tesque, Kyoto, Japan) in 4:1 acetone:olive oil (Nacalai Tesque). On day 5, both sides of the ears were challenged with 10 μL of 0.2% (vol/vol) DNFB. On day 7, ear thickness was measured with a micrometer (MDC-25MJ 293-230, Mitsutoyo, Kawasaki, Japan). To evaluate fatty acid activity, the ear skin of mice was treated topically with t10,c15-18:2, or c9,c15-18:2 (both produced from α-linoleic acid by using microbial enzymes) (Kishino et al., 2003; Tsuji et al., 2022); these compounds were dissolved in 50% (vol/vol) ethanol in phosphate-buffered saline (PBS) and provided at a dose of 1 µg/animal at 30 min before sensitization with DNFB on day 0 and before elicitation with DNFB on day 5. To assess fatty acid activity after the challenge on day 5, t10,c15-18:2 was topically applied to the ear skin on day 6. Control mice received 50% (vol/vol) ethanol in PBS as a vehicle control. In another experiment, we topically administered the retinoid X receptor (RXR) pan-antagonist HX531 (Cayman Chemical) at a dose of 40 nmol to the ear skin of mice. The HX531 was prepared in a solution containing 50% (vol/vol) dimethyl sulfoxide and 25% (vol/vol) ethanol in PBS. This application occurred 60 min before the fatty acid treatment, with 15 μL of the solution applied to both sides of the ears. Ear swelling was calculated as: (ear thickness [mm] after DNFB application on day 7) – (ear thickness [mm] before DNFB application on day 0).

Cells were isolated from ear tissue and their flow cytometric analysis was performed as described previously (Saika et al., 2021). Ears were split into dorsal and ventral skin, cut into small pieces by using scissors, and incubated in 2 mg/mL collagenase (Wako Pure Chemicals, Osaka, Japan) in RPMI 1640 medium containing 2% (vol/vol) newborn calf serum (Equitech Bio, Kerrville, Texas, USA) for 90 min at 37°C with stirring. The cell preparations were filtered through cell strainers (pore size, 100 µm; BD Biosciences, Franklin Lakes, New Jersey, USA) and then used for flow cytometric analysis.

For flow cytometric analysis, cells were suspended in 2% (vol/vol) newborn calf serum in PBS and treated with anti-CD16/32 antibody (Tru Stain fcX, BioLegend, San Diego, California, USA) to prevent nonspecific staining. The cells were washed and further stained with the following antibodies: phycoerythrin (PE)–anti-CD31 (BD Biosciences), PE–anti-c-kit (BD Biosciences), PE-Cy7–anti-F4/80 (BioLegend), fluorescein isothiocyanate (FITC)–anti-CD34 (BD Biosciences), FITC–anti-Ly6G (BioLegend), FITC–anti-CD63 (gift from Dr. Kurashima, The University of Tokyo) (Kurashima et al., 2012), allophycocyanin (APC)–anti-CD49f (BioLegend), APC–anti-Fc epsilon receptor 1 (FcϵRI, eBioscience, San Diego, California, USA), APC-Cy7–anti-CD11b (BioLegend), and brilliant violet (BV) 421–anti-CD45 (BioLegend). Dead cells were detected by using 7-aminoactinomycin D (7-AAD, BioLegend) and were excluded from analysis. Samples were analyzed MACSQuant (Miltenyi Biotec, Bergish Gladbach, Germany) or FACSAria (BD Biosciences). Cells were isolated using FACSAria through flow cytometry. Keratinocytes are gated as 7-AAD⁻ CD45⁻ CD31⁻ CD34⁻ CD49f⁺ (Saika et al., 2021), and macrophages are gated as 7-AAD⁻ CD45⁺ Ly6G⁻ F4/80⁺ CD11b⁺ cells (Nagatake et al., 2022). These cells are used for RNA extraction to assess gene expression levels. Data analysis was conducted using FlowJo 9.9 software (Tree Star, Ashland, Oregon, USA).

The assay was performed as described previously with modifications (Saika et al., 2021). Briefly, at 60 min before euthanasia on day 7 of contact hypersensitivity induction, mice were injected intravenously with 1% (wt/vol) Evans blue dye in PBS. Harvested ears were incubated in 1 M phosphoric acid (Nacalai Tesque) at 37°C overnight to extract the dye. Potassium hydroxide and acetone were added to the extract, and the resultant solution was left to phase-separate at room temperature for at least 30 min. The absorbance (OD₆₂₀) of the aqueous phase was measured in a spectrophotometer (SmartSpec Plus, Bio-Rad Laboratories, Hercules, California, USA).

Analysis was performed as described previously (Saika et al., 2020). Briefly, ear samples were embedded in Tissue-Tek OCT compound (Sakura Finetek, Osaka, Japan), frozen in liquid nitrogen and cut into sections (7 µm) by using a cryostat (CM3050 S, Leica, Wetzlar, Germany). The sections were washed with running water for 10 min, stained with Mayer hematoxylin solution (Wako) for 10 min, and washed with running water for 30 min. The sections were then stained with 1% eosin Y solution (Wako) for 1 min, washed with running water for 10 s, and dehydrated through increasing concentrations of ethanol (1 min at each concentration, 70% to 100%, Nacalai Tesque) and finally in xylene (Nacalai Tesque) for 3 min. They were mounted (Permount, Falma, Tokyo, Japan) and examined under a microscope (BZ-9000, Keyence, Osaka, Japan).

Peritoneal mast cells (PMCs) were prepared as previously reported (Meurer et al., 2016; Sawane et al., 2019). In brief, 9 mL of PBS was injected intraperitoneally into a naïve mouse by using a 20-gauge needle, the abdomen was gently massaged for 1 min to detach peritoneal cells, and then the peritoneal fluid was collected and centrifuged at 400 × g and 4°C for 5 min. The pellet was washed with RMPI 1640 medium containing 20% (vol/vol) fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin and transferred to a 10-cm dish. Cells were cultured in RMPI 1640 supplemented with 10 ng/mL IL-3 (PeproTech, Cranbury, New Jersey, USA) and 30 ng/mL stem cell factor (PeproTech) in an incubator (37°C and 5% CO₂) for 2 days, after which the supernatant and non-adherent cells were removed, and fresh culture medium was added. On day 9, the cells were collected by washing the plate three times with PBS (10 mL each time); the cell-containing washes were pooled in a 50-mL tube, which was centrifuged at 400 × g for 5 min. The pellet was recovered and moved to a fresh 10-cm dish containing RPMI 1640 supplemented with 20% fetal bovine serum, 100 U/mL penicillin, 100 µg/mL streptomycin, 10 ng/mL IL-3, and 30 ng/mL stem cell factor in an incubator (37°C and 5% CO₂) for 4 to 5 days. The efficacy of cell recovery and percentage of differentiation to PMCs were assessed by flow cytometry as the FcϵRI⁺ c-Kit⁺ CD45⁺ population; PMC populations that were more than 90% pure were used for the degranulation assay.

The assays were performed as described previously (Sawane et al., 2019) with modifications. For the IgE-dependent degranulation assay, PMCs were seeded into 96-well plates at 2 × 10⁵ cells/well, incubated for 24 h, and then sensitized with 0.2 mg/mL anti-dinitrophenyl (DNP)–IgE (Sigma-Aldrich, St. Louis, Missouri, USA) for 24 h. Cells were washed twice with Hanks’ Balanced Salt Solution (Nacalai Tesque) and stimulated with 100 ng/mL DNP–bovine serum albumin (BSA; LSL, Tokyo, Japan) for 30 min at 37°C. To assess the effect of lipid metabolites on degranulation, t10,c15-18:2 (final concentration, 300 nM) or 0.1% (vol/vol) ethanol in Hanks’ Balanced Salt Solution as a vehicle control was added to cells 30 min before stimulation with DNP–BSA. The sample size for the control group is n = 5, and for the DNP-BSA stimulated group, it is n = 7/group. For the IgE-independent degranulation assay, PMCs were seeded into 96-well plates at 2 × 10⁵ cells/well and incubated for 24 h, after which adenosine-5’-triphosphate disodium salt hydrate (ATP; final concentration, 0.1 nM) or 2,4,6-trinitrobenzene sulfonic acid (TNBS; final concentration, 1 mM) was added for 1 h. To evaluate baseline and confirm PMCs degranulation levels, we established a control group labeled “naïve,” consisting of unstimulated PMCs that did not receive ATP or TNBS treatment. t10,c15-18:2 (final concentration, 300 nM) or 0.1% (vol/vol) ethanol in PBS as a vehicle control was added to cells 30 min before stimulation with ATP or TNBS. The sample size for the non-stimulation group is n = 5/group, and the stimulated group, it is n = 4 to 6/group. Following the stimulation period, PMCs were kept on ice for 30 min, washed with PBS, and stained on ice with anti-CD63 antibody as a marker for degranulation. The degranulation level of PMCs was measured by flow cytometry.

The procedures were performed as described previously (Nagatake et al., 2018). Briefly, total RNA was isolated using Sepazol (Nacalai Tesque) from HaCaT cells or cells sorted from day 7 ear tissue, specifically keratinocytes (7-AAD⁻ CD45⁻ CD31⁻ CD34⁻ CD49f⁺) and macrophages (7-AAD⁻ CD45⁺ Ly6G⁻ F4/80⁺ CD11b⁺). RNA samples were incubated with DNase I (Thermo Fisher Scientific, Waltham, Massachusetts, USA) and reverse transcribed into cDNA by using a Super Script VIRO cDNA Synthesis Kit (Thermo Fisher Scientific). Total RNA was extracted from ear tissues by using a Relia Prep RNA Tissue Miniprep System (Promega) and reverse transcribed. Quantitative real-time PCR analysis was performed by using a LightCycler 480 II (Roche, Basel, Switzerland) and FastStart Essential DNA Probes Master (Roche). Primer sequences were: Vegfa forward, 5′-caggctgctgtaacgatgaa-3′; Vegfa reverse, 5′-gctttggtgaggtttgatcc-3′; Actb forward, 5′-aaggccaaccgtgaaaagat-3′; Actb reverse, 5′-gtggtacgaccagaggcatac-3′; VEGFA forward, 5’-tgtgtgtgtgtgagtggttga-3’; VEGFA reverse, 5’-tctctgtgcctcgggaag-3’; ACTB forward, 5’-catgtacgttgctatccaggc-3’; and ACTB reverse, 5’-ctccttaatgtcacgcacgat-3’.

The amount of VEGF-A protein in ear homogenates was analyzed by using a Mouse VEGF Quantikine ELISA Kit (R&D Systems, Minneapolis, Minnesota, USA) according to the manufacturer’s protocol. In brief, ear skin samples were homogenized for 30 s with one 4.8-φ bead and three 3.2-φ beads in PBS containing protease-inhibitor cocktail (Sigma-Aldrich) and centrifuged (9100 × g, 20 min, 4°C) as followed previous study (Nagatake et al., 2022). The supernatant was collected and diluted to a protein concentration of 4 mg/mL with PBS containing a protease inhibitor cocktail for ELISA analysis. A microplate reader (Bio-Rad Laboratories) was used to measure absorbance at OD₄₅₀.

Fatty acids were tested for their ability to activate nuclear receptors by using human RXRα, RXRβ, and RXRγ luciferase reporter assay systems (Indigo Biosciences, State College, Pennsylvania, USA) according to the manufacturer’s procedure. 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. Reporter cells were incubated with the test compounds (final concentration, 30 µM) for 24 h at 37°C in 5% CO₂. Light emission was measured in a microplate luminometer (Arvo X2, Perkin Elmer, Waltham, Massachusetts, USA), and the activities of the nuclear receptors were quantified as relative light units.

HaCaT cell culture was performed as described previously with some modifications (Saika et al., 2021). HaCaT cells (Boukamp et al., 1988) were obtained from CLS Cell Lines Service (Eppelheim, Germany) and grown in Dulbecco’s modified Eagle’s medium with high glucose (DMEM; Sigma-Aldrich) supplemented with 10% (vol/vol) FBS (Gibco), and 100 U/mL penicillin and 100 µg/mL streptomycin at 37°C and 5% CO₂. HaCaT cells were seeded in 96-well plates at 3 × 10⁴ cells/well, and cultured for 24 h. Then, the medium was replaced with DMEM without FBS and the cells were treated first with 300 nM t10,c15-18:2 for 30 min and then with 100 ng/mL recombinant human IFN-γ (PeproTech) for 24 h. We used 0.2% (vol/vol) ethanol in DMEM as vehicle control.

Lipids were extracted as previously reported (Nagatake et al., 2022). In brief, for murine serum samples (n = 6 mice/group), 50 µL of serum was added to 450 µL of methanol (Wako Pure Chemicals) and vortexed twice for 10 s each time. For fecal samples (n = 6 mice/group), fecal pellets were combined with five 5-mm zirconia beads (M&S Instruments, Osaka, Japan) in methanol and then homogenized at 6500 rpm by using the Precellys lysis and homogenization system (Bertin Instruments, Montigny-le-Bretonneux, France) twice for 15 s each time. Samples were stored overnight at –30°C for extraction. Samples were centrifuged at 1600 × g, 4°C for 10 min. The supernatant (200 µL) was mixed with a deuterium-labeled internal standard (15(S)-hydroxyeicosatetraenoic acid-d₈, Cayman Chemical) and 200 µL water (Wako Pure Chemicals) and centrifuged at 10,000 × g, 4°C for 1 min. The supernatant underwent solid-phase extraction using Sep-Pak C₁₈ cartridges (Waters, Milford, Massachusetts, USA).

LC-MS/MS was performed as reported previously (Nagatake et al., 2022). Briefly, lipids were obtained using a Monospin C₁₈-AX centrifugal column with deuterium-labeled internal standard. Fatty acid metabolites were analyzed with a Shimadzu LCMS-8050 system with a triple-quadrupole mass spectrometer (Shimadzu, Kyoto, Japan). The chromatographic separation used a Chiralcel OJ-3R column (150 × 4.6 mm, 3.0 μm; Daicel, Osaka, Japan). Solvent A was 0.1% acetic acid, solvent B was methanol, the flow rate was 0.4 mL/min, and the oven temperature was 40 °C. The metabolites were eluted with the following gradient: 10%–75% solvent B from 0−5 min, 75% solvent B for 15 min, 75%–90% solvent B from 20−25 min, 90% solvent B for 10 min, 90%–100% solvent B from 35−50 min, and 100% solvent B for 8 min, with 100%–10% solvent B for 58–59.1 min and 10% solvent B for 0.9 min for column wash and equilibration, respectively. The injection volume was 1 μL. For MS, nitrogen was used as drying gas (flow rate 10 L/min), nebulizing gas (2.5 L/min), and heating gas (10 L/min). The temperatures were set at 400°C for the heat block, 270°C for the ESI interface, and 477°C for the desolvation line. For lipidomic analysis, LC-MS raw data were preprocessed by using LabSolutions (Shimadzu) for peak alignment, noise filtering, and data extraction. Fatty acid levels were normalized as the peak area ratios of each fatty acid to the respective internal standard. Deuterated internal standards were measured to check recoveries of fatty acid metabolites. For the quantification of fatty acid metabolites, calibration curves were drawn by using fatty acid standards.

Data were analyzed by using the non-parametric Kruskal–Wallis test followed by the Dunn multiple-comparison test or the Mann–Whitney U test (Prism 6, GraphPad Software, San Diego, California, USA). A P value of less than 0.05 was considered significant.

The fatty acids t10,c15-18:2 and c9,c15-18:2 are end products of bacterial α-linolenic acid saturation metabolism ( Figure 1 ). Because increased intake of α-linolenic acid from the diet enhances the production of its derived metabolites (Nagatake et al., 2022), we initially investigated whether dietary omega-3 fatty acid intake increases the levels of t10,c15-18:2 and c9,c15-18:2. We provided mice with diets based either on linseed oil, which is high in omega-3 α-linolenic acid, or on conventional soybean oil, which is rich in omega-6 linoleic acid, and found that the levels of α-linolenic acid, t10,c15-18:2, and c9,c15-18:2 were higher in both the feces and serum from mice on the linseed oil–based diet than from mice on the soybean oil–based diet ( Figure 2A ). To assess whether the presence of these metabolites in mice relied on intestinal bacteria, we administered a linseed oil–based diet to both SPF and GF mice for 2 months. The fecal levels of α-linolenic acid were similar between GF and SPF mice, but the levels of t10,c15-18:2 and c9,c15-18:2 were higher in SPF mice than in GF mice, thus suggesting that the production of these metabolites depends on intestinal bacteria ( Figure 2B ).

We then used the mouse model of DNFB-induced allergic contact hypersensitivity to explore the immunomodulatory roles of t10,c15-18:2 and c9,c15-18:2 in this process. Pre-treatment with t10,c15-18:2 reduced ear swelling, a marker of an inflammatory condition in this model, whereas c9,c15-18:2 did not exert a similar effect ( Figure 3A ). Enhancement of vascular permeability plays a pivotal role in the development of ear swelling (Zhang et al., 2006; Yuan et al., 2010; Huggenberger and Detmar, 2011; Ono et al., 2017). To assess vascular leakage at the inflammation site, we performed an experiment involving Evans blue dye. In the DNFB-induced inflamed ears of mice treated with the vehicle only, blue dye was distributed broadly due to vascular leakage after the injection of Evans blue ( Figure 3B ). However, dye distribution was curtailed in the ears of mice treated with t10,c15-18:2, suggesting reduced leakage, whereas it remained widespread despite treatment with c9,c15-18:2 ( Figure 3B ). Quantification of Evans blue extracted from the ear tissue indicated that dye accumulation was reduced after pre-treatment with t10,c15-18:2, but c9,c15-18:2 pre-treatment did not exert this effect ( Figure 3C ). Histologic analysis revealed that—unlike c9,c15-18:2—t10,c15-18:2 inhibited the formation of epidermal edema (spongiosis), a characteristic feature of contact hypersensitivity ( Figure 3D ). These findings indicate that t10,c15-18:2 mitigated skin inflammation in mice, particularly by attenuating vascular permeability.

Vascular permeability during contact hypersensitivity is heightened via two primary mechanisms related to mast cell degranulation and VEGF-A production (Kunstfeld et al., 2004; Yamamoto et al., 2007; Hoppe et al., 2020). To evaluate whether t10,c15-18:2 inhibited mast cell degranulation, we measured mast cell expression of CD63, a known marker of mast cell degranulation (Kurashima et al., 2012; Sawane et al., 2019). Mast cells undergo degranulation in an IgE-independent fashion during contact hypersensitivity and release pro-inflammatory mediators, including histamine and proteases (Honda et al., 2013). Because this IgE-independent reaction is induced by ATP or haptens, we stimulated PMCs with ATP or TNBS, a water-soluble hapten, and evaluated subsequent degranulation (Manabe et al., 2017; Hoppe et al., 2020). Pre-treatment with t10,c15-18:2 did not diminish the expression of CD63, which increased after mast cell exposure to either ATP or TNBS ( Figures 4A, B ). We also confirmed that t10,c15-18:2 had scant effect on IgE-dependent degranulation ( Supplementary Figure 1 ), which plays a minimal role in the DNFB-induced contact hypersensitivity model in C57BL6J mice (Nagai et al., 2000).

We then focused on VEGF-A, a potent regulator of vascular endothelial cells known for its role in enhancing vascular permeability (Shibuya, 2011; Lee et al., 2021). Whereas the VEGF-A level was elevated in ear tissues treated with the vehicle only, its concentration was lower in ear skin treated with t10,c15-18:2 than in vehicle-treated samples ( Figure 4C ). Both keratinocytes and macrophages secrete VEGF-A during skin inflammation (Johnson and Wilgus, 2014); therefore, we assessed Vegfa gene expression in keratinocytes and macrophages isolated from the ear tissues. Topical application with t10,c15-18:2 decreased Vegfa expression in keratinocytes ( Figure 4D ) but not macrophages ( Figure 4E ). These results indicate that t10,c15-18:2 suppressed Vegfa expression in keratinocytes, subsequently reducing edema. In addition, we confirmed that t10,c15-18:2 inhibits VEGFA gene expression in keratinocytes using the human keratinocyte cell line, HaCaT cells ( Supplementary Figure 2 ), indicating that t10,c15-18:2 directly affects to keratinocyte function.

Furthermore, it has been reported that mice overexpressing VEGF-A in the epidermis failed to down-regulate inflammation in delayed-type hypersensitivity (Kunstfeld et al., 2004), indicating that VEGF-A is a target for reducing skin inflammation. We confirmed that topically applied with t10,c15-18:2 in post-elicitation, specifically on day 6, also reduced ear swelling ( Supplementary Figure 3 ). This supports the effect that t10,c15-18:2 is effective in reducing inflammation and swelling following its onset.

RXRs are highly expressed in keratinocytes (Saika et al., 2021); therefore, we explored whether RXRs contribute to the anti-inflammatory action of t10,c15-18:2. We administered HX531, a pan-RXR antagonist, to mice and exposed them to t10,c15-18:2. HX531 administration abolished the inhibitory effect of t10,c15-18:2 on Vegfa expression in keratinocytes ( Figure 5A ). Additionally, in contrast to the effects of t10,c15-18:2 alone, co-treatment with t10,c15-18:2 and HX531 failed to reduce ear swelling ( Figure 5B ), a characteristic DNFB-induced symptom of contact hypersensitivity, and enhanced vascular permeability ( Figure 5C ). These results indicate that t10,c15-18:2 attenuated skin inflammation in the mouse model of DNFB-induced contact hypersensitivity through an RXR-dependent pathway, by downregulating vascular permeability via the suppression of Vegfa expression in keratinocytes.

To assess the ligand activities of RXRα, RXRβ, and RXRγ in our mouse model, we used a luciferase reporter assay. Because c9,c15-18:2 failed to significantly decrease skin inflammation, we used it as a non-functional control for comparison with the ligand activity of t10,c15-18:2. Our findings showed that t10,c15-18:2 had unique RXRγ ligand activity, which was superior to that of c9,c15-18:2 ( Figure 6 ). In contrast, neither t10,c15-18:2 nor c9,c15-18:2 had ligand activity for RXRα and RXRβ ( Supplementary Figure 4 ). These results suggest the involvement of the RXRγ-mediated pathway during t10,c15-18:2-induced suppression of contact hypersensitivity.