Balb/c mice, that is S1pr2^+/+ were purchased from Jackson Lab (Bar Harbor, ME). S1pr2^-/- mice were provided by Dr. Chun’s laboratory (15). S1pr2^fl/fl mice were gifted from Dr. Hla’s laboratory at Boston Children’s Hospital (MA, USA). K14-Cre-transgenic mice were gifted from Dr. Gallo’s laboratory at the University of California San Diego (CA, USA). Epidermis-specific S1PR2 knockout mice (S1pr2^fl/fl K14-Cre) were generated by crossing S1pr2^fl/fl mice with K14-Cre-transgenic mice as described previously (16). All mice were housed in the laboratory animal facility at the University of California San Diego in individual ventilated plastic cages and were provided food and water ad libitum under temperature-controlled conditions with a 12 h light/dark cycles. All experiments were performed with 3 times with at least 3 mice in each group each time. For a total of at least 9 mice were used for each group in the experiment. All animal protocols were reviewed and approved by University of California San Diego (IACUC approval number: s10288).

Mice at 8 weeks old of S1pr2^+/+ , S1pr2^-/- , S1pr2^fl/fl , and S1pr2^fl/fl K14-Cre were depilated on the back skin 1 day before the treatment. After depilation, mice received a daily topical dose of 62.5 mg of commercially available IMQ cream (5%) (Catalog No. 51672414506, Taro Pharmaceutical Industries, Hawthorne, NY) for 7 days on the depilated back skin as described previously (17). Over 7 days, one group of mice received 10 μg of FTY720 dissolved in 80μL of ethanol 1 hour before administration of IMQ. FTY720 dosage and administration timing were per previous papers (18, 19). Skin samples and axillary lymph node (LN)s were collected for further analysis 24 hours after the last IMQ application. The severity of erythema and scale was evaluated using an objective scoring system from 0 to 4 (0, none; 1, slight; 2, moderate; 3, marked; and 4, very marked) as described previously (20). Axillary LNs were measured using Kynup Digital Caliper (Deqing Liang Feng Electronic & Technology Co., Ltd., Zhejiang, China), and weighed by Mettler Toledo Analytical Balance MS104TS/00 (Catalog No. 30133522, Mettler Toledo, Columbus, OH).

Primary normal human epidermal keratinocytes NHEKs (Catalog No. C0015C, Thermo Fisher Scientific, Waltham, MA) were cultured in EpiLife Medium with 60 mM calcium (Catalog No. MEPI500CA, Thermo Fisher Scientific, Waltham, MA) supplemented with Human Keratinocyte Growth Supplement (Catalog No. S0015, Thermo Fisher Scientific, Waltham, MA) at 37 °C in 5% CO₂. Subconfluent cells were seeded in 6-well plates and grown to 80% confluence before treatment.

NHEKs cultured in 6-well plates were incubated with phosphate-buffered saline (PBS) or 10 μM of JTE013 (Catalog No. 10009458, Cayman Chemical Company, Ann Arbor, MI) at 37°C in 5% CO₂ for 2 hours (13), then incubated with PBS for 24 hours or 100 μM of IMQ (Catalog No. 99011-78-6, InvivoGen, San Diego, CA) for 2, 5, or 24 hours at 37°C in 5% CO₂ (21, 22).

Total RNAs from mouse dorsal epidermis and NHEKs were obtained by Quick RNA Miniprep Kit (Catalog No. R1055, Zymo Research, Irvine, CA). cDNA conversion from RNA and real-time RT-qPCR were performed according to the previous report (16). The primers and the probes used for real-time RT-qPCR are listed in Supplementary Tables S1, S2 , respectively. The volume of RNA and cDNA was measured by NanoDrop 2000 Spectrophotometer (Catalog No. ND2000CLAPTOP, Thermo Scientific, Waltham, MA). The expression of target genes was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression and analyzed by the 2^{– ΔΔct} method.

Formalin-fixed and paraffin-embedded (FFPE) mouse skin samples were sectioned, and HE stained at the University of California San Diego Biorepository and Tissue and Technology Shared Resource. For immunofluorescence staining of FFPE sections, unstained sections were deparaffinized, rehydrated, and allowed to reach boiling temperature and maintained at a sub-boiling temperature for 10 minutes with antigen retrieval solution (Catalog No. 00495558, Invitrogen, Waltham, MA). After antigen retrieval, immunofluorescent staining was performed using the primary and fluorochrome-conjugated secondary antibodies listed in Supplementary Table S3 . Nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI) and slides were mounted with Prolong Gold Anti-Fade Reagent with DAPI (Catalog No. 8961, Cell Signaling Technology, Danvers, MA).

For immunofluorescence staining of NHEKs, cells were cultured on the cover glass in 10-cm petri dish for 24 hours and fixed with 4% paraformaldehyde. Fixed cells were stained with primary and fluorochrome-conjugated secondary antibody listed in Supplementary Table S3 . For negative control, fixed cells were stained only with fluorochrome-conjugated secondary antibody. Nuclei were counterstained with DAPI with Prolong Gold Anti-Fade Reagent with DAPI (Catalog No. 8961, Cell Signaling Technology, Danvers, MA). Fluorescence images were obtained with a fluorescence microscope EVOS M5000 (Catalog No. AMF5000, Invitrogen, Waltham, MA).

Axillary LNs were extracted from mice, and LN cells were isolated for flow cytometry analysis. Surface staining was performed for 20 minutes at 4°C with fluorochrome-conjugated antibodies. Fc receptors were blocked with BD Pharmingen Purified Rat Anti-Mouse CD16/CD32 (Mouse BD Fc Block) (Catalog No. AB_394656, BD Biosciences, Franklin Lakes, NJ) diluted 1:100 before staining. For intracellular staining, the cells were fixed and permeabilized using BioLegend Cyto-Fast Fix/Perm Buffer Set (Catalog No. 426803, BioLegend, San Diego, CA), then stained with fluorochrome-conjugated antibodies at room temperature for 30 minutes according to the manufacturer’s instruction. Fixable Viability Dye eFluor 506 (Catalog No. 65-0866-14, Invitrogen, Waltham, MA) were used for viability dye. Fluorescein isothiocyanate (FITC)-conjugated anti-mouse IL-17A monoclonal antibody (mAb) (Catalog No. 506907), APC-conjugated anti-mouse CD4 mAb (Catalog No. 100515), and BV605-conjugated anti-mouse CD8 mAb (Catalog No. 100744) were purchased from BioLegend (San Diego, CA). Samples were measured in flow cytometer, and data were analyzed with FlowJo software (FlowJo, Ashland, OR).

ELISA kit (Catalog No. MBS2708370, MyBioSource, San Diego, CA) were used to determine NF-κB protein expression in mouse skin. Proteins were detected from supernatants of tissue homogenate according to the manufacturer’s instructions. Optical density (OD) value was detected with a micro-plate reader SpectraMax iD3 (Catalog No. ID3-STD, Molecular Devices, San Jose, CA). ELISA measurements were normalized to each sample’s total weight or volume as described previously (23).

FFPE mouse skin samples that passed the RNA quality control (DV200 > 30%) were used for spatial transcriptomic library construction and sequencing. The tissues were prepared according to the Visium CytAssist Spatial Gene Expression for FFPE-Tissue Preparation Guide (CG000518, 10× Genomics, Pleasanton, CA, USA). FFPE mouse skin samples were sectioned into 10 μm, then deparaffinization, probe hybridization and probe ligation were performed. The tissue was permeabilized and the mRNA was reverse transcribed into cDNA with barcode containing slide location information on a designated area of 11 mm × 11 mm square. cDNA was then amplified and processed to obtain cDNA libraries according to the manufacturer’s protocol. To verify the size of PCR-enriched fragments, the template size distribution was checked using Agilent D1000 ScreenTape (Agilent Technologies, Santa Clara, CA). Sequencing was performed at the IGM Genomics Center at the University of California, San Diego. Libraries were sequenced using NovaSeq X Plus 10B (Illumina, San Diego, CA) with a read length of 150 – 1000 bp. Partek flow software version 11 (Partek Incorporated, Chesterfield, MO, USA) was used to process H&E images and FASTQ files and perform spatial gene expression analysis and pathway enrichment analysis. The sequencing results were guaranteed to be accurate as follows; Number of reads: IMQ-treated S1pr2^fl/fl , 577,665,006; IMQ-treated S1pr2^fl/fl K14-Cre, 496,579,139; Number of spots under tissue: IMQ-treated S1pr2^fl/fl , 1,213; IMQ-treated S1pr2^fl/fl K14-Cre, 975; Mean reads per spot: 903,639; 1,010,791; Median genes per spot: 1,758; 1,046.

In all experiments, all samples were performed in triplicates, and values were expressed as mean ± SD. Two-way ANOVA and Tukey’s multiple comparison test determined statistical significances of differences for multiple group analysis. Student t-test was performed to compare two groups. P < 0.05 was considered significant. Analyses were performed through the GraphPad Prism software version 7 (GraphPad Software, Boston, MA).

To investigate whether S1PR2 has any effects on psoriasis inflammation, we induced psoriasis-like skin inflammation with 5% IMQ cream on the depilated dorsal skin of S1pr2^-/- mice, which has no S1PR2 in their whole body and that of S1pr2^+/+ mice, S1pr2 wild type control. After IMQ application, S1pr2^-/- mice had much more scales on their dorsal skin than S1pr2^+/+ mice, though the severity of erythema was not different between S1pr2^-/- mice and S1pr2^+/+ mice ( Figures 1A, D ). None developed systemic symptoms other than skin manifestations. No difference in body shape and size of the mice was observed between the S1pr2^+/+ mice and S1pr2^-/- mice. Histological analyses with hematoxylin and eosin (H&E) staining of S1pr2^-/- mice showed more parakeratosis in the stratum corneum, and more inflammatory cell infiltration in the lower dermis than that of S1pr2^+/+ mice after IMQ application ( Figure 1B ). In addition, immunofluorescence staining revealed that S1pr2^-/- mice had more neutrophils in the lower dermis than S1pr2^+/+ mice after IMQ application ( Figure 1C ). We then quantified the mRNA expression levels of proinflammatory cytokines in the epidermis. We found that S1pr2^-/- mice epidermis had significantly higher mRNA expressions of Tnf-α, Il-17a, and Il-1β than S1pr2^+/+ mice after IMQ application ( Figure 1E ). These results demonstrate the involvement of S1PR2 in psoriasis pathogenesis and suggest that deletion of S1PR2 amplifies psoriasis inflammation.

Psoriasis can be triggered by damaging stimuli such as trauma or scratching (24). Keratinocytes, which constitute the outermost layer of the skin, receive these stimuli first and directly. To assess whether the deletion of S1PR2 only in keratinocytes affects psoriasis inflammation, we induced psoriasis-like skin inflammation on the depilated dorsal skin of S1pr2^fl/fl K14-Cre mice, that lack of S1PR2 only in keratinocytes, and that of S1pr2^fl/fl mice. Consistent with the results of S1pr2^-/- mice, S1pr2^fl/fl K14-Cre mice showed more scales on their dorsal skin than S1pr2^fl/fl mice after IMQ application, though the severity of erythema was not different between S1pr2^fl/fl K14-Cre mice and S1pr2^fl/fl mice ( Figures 2A, D ). None developed systemic symptoms other than skin manifestations. No difference in body shape and size of the mice was observed between the S1pr2^fl/fl mice and S1pr2^fl/fl K14-Cre mice. H&E staining of S1pr2^fl/fl K14-Cre mice presented more parakeratosis in the stratum corneum and more inflammatory cell infiltration in the lower dermis than that of S1pr2^fl/fl mice ( Figure 2B ). Immunofluorescence staining analysis showed that S1pr2^fl/fl K14-Cre mice had more neutrophils in the lower dermis than S1pr2^fl/fl mice ( Figure 2C ). We investigated proinflammatory cytokine profiles and found that S1pr2^fl/fl K14-Cre mice epidermis had significantly higher mRNA expressions of TNF-α, Il-17A, and IL-1β than that of S1pr2^fl/fl mice after IMQ application ( Figure 2E ).

Several studies on mice revealed that the S1PR modulator fingolimod (FTY720) binds S1PR1, 3, 4, and 5, and selective S1PR1 modulator Syl930 ameliorates psoriasis inflammation (12, 25). We, therefore, investigated whether all other S1PRs could regulate inflammation in psoriasis exacerbated by S1PR2 deficiency in keratinocytes. We treated IMQ-induced psoriasis skin of S1pr2^fl/fl and S1pr2^fl/fl K14-Cre mice with a topical application of FTY720 and compared cytokine profiles with control and IMQ-only groups ( Supplementary Figure S1A ). With FTY720 and IMQ, S1pr2^fl/fl K14-Cre mice still had more scaly erythema than S1pr2^fl/fl mice ( Supplementary Figures S1B, C ), and S1pr2^fl/fl K14-Cre mice epidermis had significantly higher mRNA expressions of TNF-α and IL-1β than that of S1pr2^fl/fl mice. These results suggest that S1PR2 in keratinocytes plays a crucial role in regulating inflammation in psoriasis and that S1PR2 is more effective in controlling psoriasis than other S1PRs.

IL-17A plays a critically important role in psoriasis inflammation, and Th17 cells are thought to be the main source of IL-17A in psoriasis (26). To investigate whether S1PR2 deletion in keratinocytes affects Th17 cell increase in psoriasis-like skin, we examined the distribution of CD4⁺ IL-17A⁺ and CD8⁺ IL-17A⁺ T cells in the skin-draining LNs from S1pr2^fl/fl K14-Cre and S1pr2^fl/fl mice after IMQ application. In terms of size and weight of LNs, S1pr2^fl/fl K14-Cre mice had larger and heavier LNs than S1pr2^fl/fl mice ( Figures 3A, B ). In addition, we observed a larger number of CD4⁺IL-17⁺ T cells in the skin-draining LNs of S1pr2^fl/fl K14-Cre mice than that of S1pr2^fl/fl mice after IMQ administration, while no change was observed in the population of CD8⁺IL-17⁺ T cells between S1pr2^fl/fl K14-Cre and S1pr2^fl/fl mice ( Figures 3C–G ). We also examined the distribution of Cd4 ⁺ Il-17a ⁺ and Cd8a ⁺ Il-17a ⁺ areas in the skin by spatial sequencing to investigate the population of Th17 cells and Tc17 cells in the skin. Contrary to the LN results, a lower percentage of Cd4+Il-17a+ areas and a lower percentage of Cd8a ⁺ Il-17a ⁺ areas were observed in the IMQ-treated S1pr2^fl/fl K14-Cre mouse skin, compared to S1pr2^fl/fl mouse skin ( Supplementary Figures S2A, B ). Next, we examined Cd2+Cd3+Il-17a+ areas to investigate the distribution of γδ T cells and observed a smaller number in the IMQ-treated S1pr2^fl/fl K14-Cre mouse skin ( Supplementary Figure S2C ). It has been reported that IL-17A production by CD8+ T cells is upregulated in psoriatic skin (27), and the results of this study suggest that S1PR2 deficiency increases IL-17A production from CD8+ cells in the psoriatic skin region. These results suggest that deleting S1pr2 in mouse keratinocytes affects skin-resident cells and the recruitment of IL-17-producing cells during psoriasis-like inflammation.

The potential involvement of skin barrier dysfunction in the pathogenesis of psoriasis has been previously discussed (28). Some reports showed an increase of tight junction proteins and filaggrin in psoriasis lesions (29, 30), while others showed a decrease (31, 32). Thus, a consensus has yet to be reached. Since our previous studies have shown that S1PR2 is essential for maintaining and recovering epidermal skin barrier function (13, 16), we investigated whether the skin barrier is altered in psoriasis-like inflammation lacking epidermal S1PR2. RT- qPCR analysis showed no specific trend in mRNA expression levels of the skin barrier proteins between the control and the IMQ-treated group in either S1pr2^fl/fl or S1pr2^fl/fl K14-Cre mice ( Supplementary Figure S3 ). Our results suggest that impairment of the skin barrier function is not involved in the pathogenesis of psoriasis.

In a mouse model of psoriasis, IMQ acts as a TLR7 ligand to induce the expression of cytokines required for activation of the Th17 pathway and inflammatory factors necessary for psoriasis development (33). To explore how S1PR2 affects IMQ-induced psoriasis-like inflammation, we examined mRNA expressions of the TLR7 signaling pathway in the epidermis of S1PR2-deficient mice. In the epidermis of IMQ-treated S1pr2^-/- mice, mRNA levels of Myd88, Irf7, Nfkb1, and Rela were higher than in wild-type mice, but no significant difference was observed in the mRNA level of Tlr7 ( Figure 4A ). Similar to that, the epidermis of IMQ-treated S1pr2 ^fl/fl K14-Cre mice had higher mRNA levels of Myd88, Traf6, Irf7, Nfkb1, and Rela compared to S1pr2 ^fl/fl mice ( Figure 4B ). In addition, the epidermis of IMQ-treated S1pr2 ^fl/fl K14-Cre mice had higher protein levels of NF-κB ( Figure 4C ). This result shows that the deletion of S1PR2 causes an excessive increase in the MyD88/NF-κB pathway, not solely due to TLR7 stimulation.

To confirm histology and qPCR analysis, we performed spatial transcriptomics of whole skin tissue from IMQ-treated S1pr2 ^fl/fl and S1pr2 ^fl/fl K14-Cre mouse. Graph-based clustering analysis and Uniform Manifold Approximation and Projection for Dimension Reduction (UMAP) showed that epidermis and dermis were classified in separate clusters for IMQ-treated S1pr2 ^fl/fl and S1pr2 ^fl/fl K14-Cre mouse, although subcutaneous tissue was classified as a same cluster in both IMQ-treated S1pr2 ^fl/fl and S1pr2 ^fl/fl K14-Cre mouse ( Figure 5A ). In addition, the spatial gene expression analysis showed an upregulation of Tnf, Il-17a, Il-17c, Il-17f, and Il-1β in the epidermis of IMQ-treated S1pr2 ^fl/fl K14-Cre mouse, consistent with the findings from the pathway enrichment analysis ( Figures 5B, C ). In addition, gene expression levels of Il-12 cytokine families such as Il-12a, Il-12b, Il-23a, and Il-27 were upregulated in the epidermis of IMQ-treated S1pr2 ^fl/fl K14-Cre mouse ( Figures 5B, C ). These data support the fact that the IL-23/Th17 axis is recognized as a key link in the immunopathogenesis of psoriasis, and S1P has been reported to affect IL-23-mediated signaling (34). Moreover, the spatial gene expression analysis and pathway enrichment analysis revealed that the MyD88/NF-κB pathway is upregulated in IMQ-treated S1pr2 ^fl/fl K14-Cre mouse epidermis, while TLR domain-containing adapter inducing IFN-β (TRIF) and TRAF3, a MyD88-independent signaling pathway, is downregulated ( Figure 5C , lower panel). These results support the histology and qPCR findings that S1PR2 deletion in keratinocytes induces MyD88/NF-κB pathway increase in the epidermis, resulting in proinflammatory cytokine increase after IMQ stimulation. Since angiogenesis is an exacerbating factor of psoriatic inflammation (35), we also investigated the gene expression levels of vascular endothelial growth factor (VEGF), hypoxia-inducible factor (HIF), the murine IL-8 homolog CXCL1, and angiopoietin that are considered to be the main players in angiogenesis ( Supplementary Figure S4 ). The spatial gene expression analysis and pathway enrichment analysis showed that S1PR2 deletion did not alter these expressions or even reduce them. This is consistent with the fact that S1PR2 does not change the severity of erythema, suggesting that epidermal S1PR2 does not affect angiogenesis.