Skin γδ T cell inflammatory responses are hardwired in the thymus by oxysterol sensing via GPR183 and calibrated by dietary cholesterol

Summary

Dietary components and metabolites have a profound impact on immunity and inflammation. Here, we investigated how sensing of cholesterol metabolite oxysterols by γδ T cells impacts their tissue residency and function. We show that dermal IL-17 producing γδ T (Tγδ17) cells essential for skin barrier homeostasis require oxysterols sensing through G protein receptor 183 (GPR183) for their development and inflammatory responses. Single-cell transcriptomics and murine reporter strains revealed that GPR183 on developing γδ thymocytes is needed for their maturation by sensing medullary thymic epithelial cell-derived oxysterols. In the skin, basal keratinocytes expressing the oxysterol enzyme cholesterol 25-hydroxylase (CH25H) maintain dermal Tγδ17 cells. Diet-driven increases in oxysterols exacerbate Tγδ17 cell-mediated psoriatic inflammation, dependent on GPR183 on γδ T cells. Hence, cholesterol-derived oxysterols control spatially distinct but biologically linked processes of thymic education and peripheral function of dermal T cells, implicating diet as a focal parameter of dermal Tγδ17 cells.

Introduction

Immunocytes in mucocutaneous barrier tissues have garnered increasing prominence as the first responders to infections, while also contributing to tissue homeostasis and organogenesis.¹ The latter facet of tissue-resident lymphocytes is especially critical in early animal development and abnormalities have been proposed to contribute to child developmental defects ranging from behavioral alterations to immune hypersensitivity in diverse tissues.^2–4 The skin, lung and gut are replete with non-conventional T cells: innate T cells and innate lymphoid cells (ILCs) that are preprogrammed for effector function preferentially populate barrier tissues and confer immediacy in responses to perturbations. This specialization contrasts to adaptive lymphocytes that differentiate into effector subsets days after encounter with pathogens. Moreover, to perform homeostatic maintenance in tissues, the mucosal immune system often senses non-pathogen derived products that reflect tissue health and metabolic states.^5,6

Tissue-resident Type 3 cytokine (IL-17, IL-22) producing T cells are prototypic innate T cells that serve critical function in skin, lung, and adipose tissue homeostasis. One prominent T3L subset in mice is IL-17 producing γδ T cells (Vγ2TCR⁺ Tγδ17, Raulet nomenclature⁷) that originate from fetal progenitors and populate the dermis soon after birth.⁸ These neonatal Tγδ17 cells are essential to prevent adult-onset atopic dermatitis,⁹ but are also critical for driving inflammation in a widely utilized mouse model of psoriasis.¹⁰ How Tγδ17 cells, and other T3L (CD4⁺ Th17, ILC3, MAIT17, NKT17), populate specific non-lymphoid tissues to mediate context-dependent function remains unclear. Possible drivers of anatomical specificity of T3L are the two G protein-coupled receptors (GPCRs) CCR6 and GPR183 (also known as Ebi2) expressed on nearly all T3L in rodents¹¹ and humans.¹² While CCR6-CCL20 function in the extravasation of T cells from circulation is well established, Ccr6^−/− Tγδ17 cells do not exhibit significant deficits in effector function in mouse models of psoriasis.¹³ GPR183 is critical for proper localization of many immune cells, including: tissue positioning of follicular B cells,¹⁴ Tfh clustering with activated dendritic cells,¹⁵ naive CD4 T cells intranodal localization,¹⁶ and ILC3-mediated colonic lymphoid microstructure formations.¹¹ The ligand for GPR183 is the oxysterol 7α,25-hydroxycholesterol (HC) and its production is controlled by cholesterol 25-hydroxylase (CH25H).¹⁷ Cholesterol homeostasis impacts production of two cytokines that drive the activation of T3L, IL-23^18,19 and IL-1, in dendritic cells (DCs) and macrophages,²⁰ respectively.

Here we demonstrated that Tγδ17 cell development and skin localization proceed by sensing anatomical depots of oxysterols via GPR183. In the thymus, we showed that a specific subset of Sox4⁺Ch25h⁺ medullary thymic epithelial cells (mTECs) enforced maturation checkpoints to permit Tγδ17 cell homing to the dermis. Severity of skin inflammatory responses in humans is linked to diet with high fat content.³ In the skin, Tγδ17 cell-mediated psoriasis induction was regulated by Ch25h⁺ cells. Mice fed high cholesterol food experienced more severe psoriasis and this disease progression required GPR183-sufficient Tγδ17 cells. Hence, spatially discreet dietary cholesterol metabolites are the major determinant of skin innate T3L generation and function. The GRP183-oxysterol pathway emerges as the candidate mechanism underpinning high fat diet-induced aberrant tissue inflammation in humans.

Results

CCR6 and GPR183 are co-expressed on mature Tγδ17 thymocytes and required for their development.

To identify the earliest gene circuits that specify mucocutaneous tissue homing of γδ T cells, we performed single cell RNA-sequencing (scRNA-seq) analyses on developing γδTCR⁺ thymocytes. Unsupervised principal-component analysis (PCA), and t-stochastic neighbor embedding (tSNE) identified 7 clusters, to generate the three distinct effector subsets, Tγδ17, Type 1 cytokine secreting γδ T cells (Tγδ1) and dual IFNγ and IL-4 secreting γδTCR⁺ NKT cells (γδNKT) (Figures 1A and S1A). Most γδ thymocytes expressed Cd24 and Sox13 High Mobility Group transcription factor (HMG TF), markers of immaturity and γδ T cell lineage, respectively. At the immature stage, two bifurcating effectors subsets marked by RORγt and LEF1 were discerned (Figure 1B). RORγt regulates Il17 transcription and LEF1 HMG TF is linked to Tγδ1 programming.^10,21 Similar to other T3L, mature Tγδ17 cells uniquely express Ccr6 and Gpr183 (Ebi2), with the expression of Gpr183 detected in some immature γδ thymocytes. Using a Gpr183 transcriptional reporter (Gpr183^Egfp) model we verified Gpr183 and CCR6 co-expression patterns in mature Tγδ17 cells (Figures 1C, 1D). There are two Tγδ17 cell subtypes in mice: Vγ4⁺ (Vγ6 by the Tonegawa nomenclature) T cells that arise exclusively during fetal development and are functionally critical in several tissues,²² and neonatal-origin Vγ2⁺ T cells with non-redundant function in the skin and lung.⁹ Gpr183 and CCR6 coexpression was only observed in mature (CD24^neg) Vγ2⁺ CD27^neg thymocytes, and to a lesser extent, mature Vγ4⁺ (identified here as cells that do not stain with Ab to Vγ1.1 and Vγ2TCR, the two dominant subtypes in the thymus) cells, at all ages (Figures 1C and S1B). Expression of the scavenger receptor SCART2 is skewed to Vγ2⁺ Tγδ17 cells²³ and nearly all mature SCART2⁺ (CD27^neg) thymocytes coexpressed Gpr183 and CCR6. Among αβ T cell lineage thymocytes, there was virtually no CCR6 expression, and most did not transcribe Gpr183, with only the mature CD4⁺ thymocytes being the exception (Figure S1C). In the skin, nearly all dermal Vγ2⁺, and majority of Vγ4⁺, Tγδ17 cells co-expressed Gpr183 and CCR6, while Vγ3⁺ dendritic epidermal T cells (DETCs) did not express either marker (Figures 1D and S1D). None of dermal αβTCR⁺ T3L²⁴ showed the homogeneous dual Gpr183 and CCR6 expression pattern (Figure S1D).

Figure 1. Intrathymic programming of Ccr6 and Gpr183 coexpression on mature Tγδ17 thymocytes.

(A) Two-dimensional tSNE graphical representation of 3,749 single TCRδ⁺ thymocytes sorted from 12 day old mice (n=7), clustered into seven color-coded TCRδ⁺ developmental intermediates.

(B)Feature plots of distribution of cells expressing select signature genes of Tγδ17 and Tγδ1 subtypes, superimposed onto the two-dimensional graphs.

(C)Representative flow cytometric contour plots depicting expression of Gpr183, CCR6, and SCART2 on immature (CD24⁺, blue) and mature (CD24⁻, red) Tγδ17 thymocytes (CD27^neg at the mature stage) from Gpr183 reporter mice.

(D) Representative flow cytometric analysis of expression amounts of Gpr183 and CCR6 of skin γδ T subsets distinguished by TCRγ chain (DETCs express Vγ3TCR) from Gpr183 reporter mice.

Data in (C-D) are representative of at least three independent experiments.

CCR6 contributes to immunocyte migration to skin²⁵ but it is reported to be not essential for skin γδ T cells.^13,26 Function of GPR183 on skin T cells is unknown. We generated Gpr183^−/−Ccr6^−/− double gene knock-out (DKO) mice to determine contributions of each GPCR to dermal T cell localization. In the skin, there were no significant differences in total lymphoid cellularity, and γδ DETCs and αβ T cell subset distributions in all genotype combinations assessed were unaltered (Figures 2 and S2A). In sharp contrast, neonatal-origin Tγδ17 cells were severely depleted in the skin of DKO mice (Figures 2A, 2B and 2C). While compound heterozygotes had diminished or comparable neonatal Tγδ17 cell frequency relative to Gpr183^/- or Ccr6^−/− mice, respectively (Figures S2B and S2C), only double deficiency virtually abolished these cells in the skin. That the loss of dermal Vγ2⁺ Tγδ17 cells was not primarily caused by a block in transit from LNs to the skin was suggested by the specific paucity of IL-17⁺ Vγ2⁺ T cells already in skin draining LNs (sLNs) of DKO mice (Figures 2D, 2E and 2F). Moreover, the defect in Vγ2⁺ Tγδ17 cell maturation was precociously evident in the thymus of DKO mice, where mature, but not immature, Vγ2⁺ Tγδ17 thymocytes were significantly depleted (Figures 2G and 2H), while all other thymocyte subsets were phenotypically and numerically in the normal range (Figure S2D), as previously shown also for Gpr183^−/− or Ccr6^−/− mice.²⁷ The maturation-deficit in Tγδ17 cells was not observed in compound heterozygotes (Figure S2E). Thus, while either GPR183 or CCR6 is sufficient for Vγ2⁺ Tγδ17 cell production, their concomitant absence is not permissive for Tγδ17 cell terminal maturation in the thymus.

Figure 2. CCR6 and GPR183 are required for dermal Vγ2⁺ Tγδ17 cell production in the thymus.

(A) Representative flow cytometric plots of γδTCR⁺ subsets (left) and intracellular IL-17 (right) expression amongst γδ T cells in the skin of adult Ccr6^+/−Gpr183^+/− compound heterozygotes and Ccr6^−/−Gpr183^−/− DKO mice. TCRδ^hiVγ2TCR^neg cells are DETCs.

(B) Summary of the compiled flow cytometric data of A (n≥11 mice per genotype). P values determined by unpaired t-test.

(C) Pie charts showing the frequencies of different T cell subsets making up IL-17⁺ skin cells of compound heterozygotes and DKO mice analyzed in A.

(D-F) Corresponding sets of data from sLN.

(G) Representative flow cytometric plots of γδTCR⁺ thymic subsets (left; Vγ2⁻Vγ1.1⁻ cells are mostly Vγ4TCR⁺) and Vγ2TCR⁺ thymocytes distinguished by CD27 and SCART2 expression (right; red box indicates mature Tγδ17 thymocytes) in adult compound heterozygotes and DKO mice.

(H) Summary of the compiled flow cytometric data as in G (n≥11 mice per genotype).i

Data in (A-H) are representative of at least three independent experiments.

In addition to 7α,25-HC whose production relies on CH25H, GPR183 can also recognize 7α,27-HC (Figure S2F), an oxysterol produced by the enzyme CYP27A1,¹⁷ albeit with lower potency. Therefore, to test whether CH25H plays a dominant role in producing the GPR183 ligand required for Tγδ17 cells, we generated mice lacking both CCR6 and CH25H. In these animals, Vγ2⁺ Tγδ17 cells were depleted in the skin and sLNs and phenocopied Ccr6^−/−Gpr183^−/− DKO mice (Figures S2G and S2H). In the thymus of Ccr6^−/−Ch25h^−/− mice, mature Vγ2⁺ Tγδ17 cell production was significantly impaired, but not blocked (Figure S2I). These results suggest that while CH25H is the dominant enzyme involved in the generation of GPR183 ligand that assures proper maintenance of Tγδ17 cells extrathymically, CYP27A1 may partly compensate for the loss of CH25H in the thymus.

A subset of medullary thymic epithelial cells expresses Ch25h to position Tγδ17 thymocytes

The block in the generation of mature Tγδ17 thymocytes in DKO mice unexpectedly suggests that coordinated sensing of oxysterols and CCL20 in the thymus is a prerequisite for proper maturation and egress to peripheral tissues. Normal thymic epithelial cells (TECs) are required for Vγ2⁺ Tγδ17 cell development.^28,29 While CCL20 is produced by several types of mTECs and oxysterols are detectable in thymus extracts,³⁰ the cellular source of thymic GPR183 ligand is unknown. To establish the spatial organization of oxysterol production during Tγδ17 development, we generated a tdTom reporter mouse line that tracks the transcriptional activity of Ch25h, encoding for the key enzyme to generate 7α,25-HC.

In Ch25h-tdTom mice the reporter expressing cells were restricted to the medulla, with virtually no expression in the cortex (Figure 3A). This pattern was observed in all ages tested (Figure 3B). Aire expression (marking mTEC2, nomenclature based on³¹) was excluded from Ch25h⁺ cells (Figure 3B), while Ccl21 (CCR7 ligand) expression was coincident (Figure 3C). In comparison, Ccl20 expression was widespread in mTECs (Figure 3C). We extended these findings with scRNAseq analyses of thymic epithelial subsets (Figures 3D and S3). There exists a unique cluster of Ch25h expressing mTECs that belonged to Ccl21⁺ mTEC1.³² Some of these Ch25h⁺ mTECs also expressed Cyp7b1 and other enzymes in the sterol synthesis pathway, but not Hsd3b7, a dehydrogenase that inactivates GPR183 ligand to generate bile acid precursors, as illustrated by a transcriptome heatmap (Figure 3D) and feature plots of expression of select signature genes of TEC subsets and sterol biosynthetic pathway (Figures S3A, S3B, S3C and S3D), with additional verification of expression of key genes by RT-qPCR (Figure S3E). Ch25h was not detected in Aire⁺ mTEC2 and Il25⁺ mTEC4,^31,33 which mediate tolerance induction to tissue antigens and promote the development of Type 2 cytokine producing innate-like thymocytes, respectively (Figures 3D, S3A, S3B and S3C). Human thymus sc transcriptome public database indicates a homologous Ch25h expression pattern in human TECs,³⁴ suggesting that the thymic oxysterol niche is conserved in mammalian evolution.

Figure 3. Ch25h⁺ mTEC1 constitutes the oxysterol thymic niche for Tγδ17 cell maturation.

(A) Representative immunofluorescence (IF) image (from 5 experiments) of adult Ch25h reporter thymus stained with anti-RFP Ab (red, capturing tdTom expression). M, medulla; C, cortex. Dashed white line demarcates cortico-medullary junction. Scale bar is 20 μm throughout.

(B-C) Representative IF images (from 5 experiments) of adult and neonatal Ch25h reporter thymi stained with anti-RFP (red) and anti-AIRE (green) Abs and corresponding zoom-in details (B); with anti-CCL21 Ab (green, top row) and with anti-CCL20 Ab (green, bottom row) (C).

(D) Heatmap showing differential single cTEC and mTEC expression of select signature genes of TEC subsets and genes in the sterol metabolic pathway. Color bars (top) represent identities of five distinct TEC subsets.

(E) Representative IF image (from 5 experiments) of Ch25h reporter thymus stained with anti-RFP (red) and anti-SCART2 (blue) Abs. The distance of SCART2-expressing cells from Ch25h-expressing cells was measured using ImageJ, and percentage of SCART2-expressing cells between 0–25, 25–50, and >50μM distance was plotted (≥350 cells analyzed; P values determined by unpaired t-test).

(F) Representative IF images (from 3 experiments) of SCART2-expressing cells in adult WT and DKO thymi stained with anti-SCART2 Ab (blue). Frequencies of SCART2-expressing cells localized in the medulla or the cortex are recorded for both genotypes (≥300 cells per genotype analyzed; P values determined by unpaired t-test).

Data in (A-C and E-F) are representative of at least 3 independent experiments.

To visualize the homeostatic positioning of Tγδ17 cells in the thymus, we stained thymi from Ch25h-tdTom mice with Abs to γδTCR and the unique marker of Vγ2⁺ Tγδ17 cells SCART2. Maturing SCART2⁺ Tγδ17 cells were seen at the cortical medullary junction and in the medulla, with ~90% of SCART2⁺ cells within 50um of Ch25h⁺ mTECs (Figure 3E). We next analyzed the positioning of Vγ2⁺ Tγδ17 cells in the thymus of DKO mice. While Tγδ17 cells in the thymi of WT mice were primarily observed at the cortical medullary junction and in the medulla, in DKO thymi SCART2⁺ Tγδ17 cells were mainly localized in the cortex, at considerable distance from the cortical medullary junction (Figure 3F). These results demonstrates that the thymus contains epithelial depots of oxysterols, required for normal positioning of developing Tγδ17 cells at the medulla. The coincident inability to sense 7α,25-HC and CCL20 disrupts Tγδ17 positioning and maturation.

Sox4 in TEC controls Tγδ17 thymocyte development

To determine whether Ch25h⁺ mTECs are specifically required for Tγδ17 cell development necessitated the identification of a factor regulating their development or maintenance. Survey of all significantly expressed TFs (570) in the TEC scRNAseq datasets identified SOX4,^10,21 a TF involved in various lymphocyte developmental processes, with well-defined function in innate T3L development, as enriched in Ch25h⁺ mTEC (Figures 3D and S3F). To determine whether Ch25h-expressing cells were also expressing Sox4, we generated double reporter mice (Ch25h^tdTom;Sox4^Egfp). Analysis of thymus revealed that all the Ch25h⁺ mTEC also expressed Gfp, with minimal Gfp expression in other TEC subsets (Figure 4A), including Aire⁺ mTEC2, in line with the scRNAseq data.

Figure 4. Sox4 expressing TECs control Tγδ17 thymic development.

(A) Representative IF images (from 3 experiments) of adult Sox4 reporter thymus stained with stained with anti-GFP (green, Sox4 expression proxy) and anti-AIRE (red) Abs (left image), juxtaposed to a representative image (from 2 experiments) of adult Sox4-Ch25h double reporter thymus stained with anti-GFP (green) and anti-RFP (red, Ch25h expression proxy) Abs (right image). Dashed white line demarcates cortico-medullary junction. Scale bar is 20 μm throughout.

(B) Representative IF images of adult Foxn1^creSox4^+/+ (TEC^Wt), Foxn1^creSox4^fl/+ and Foxn1^creSox4^fl/fl (TEC^ΔSox4) thymi stained with anti-KRT5 (red), anti-CD11c (blue) and anti-AIRE (green) Abs.

(C) Representative flow cytometric plots of CCR6 and CD27 expression on mature Vγ2TCR⁺ thymocytes from 10-day old TEC^WT and TEC^ΔSox4 thymi (top row). Summary of the compiled frequencies of Vγ2TCR⁺ cells amongst total γδTCR⁺ thymocytes and Vγ2⁺ Tγδ17 cells amongst total mature Vγ2TCR⁺ thymocytes in mice of indicated genotypes (middle row). Corresponding summary for Vγ4TCR⁺ cell frequencies (bottom row) with n≥5 mice per genotype.

(D) Representative intracellular staining for IL-17 amongst Vγ2TCR⁺ cells from adult TEC^Wt and TEC^ΔSox4 cervical LNs (top row). Summary of the compiled frequencies of activated (CD44⁺) cells amongst Vγ2TCR⁺ cells and Vγ2⁺IL17⁺ cells amongst total γδ T cells (middle row). All thymus-programmed Tγδ17 cells are CD44⁺ in LNs. Corresponding summary for Vγ4TCR⁺ cell frequencies (bottom row). n≥6 mice per genotype.

(E) Representative flow plots of γδ T cell subsets from adult TEC^WT and TEC^ΔSox4 ear skin (top row). Summary of the compiled frequencies of dermal Vγ2TCR⁺ and Vγ4TCR⁺ cells in the skin (bottom row) with n≥6 mice per genotype.

(F) Schematic for GPR183-mediated transwell migration assay.

(G) Relative migration efficiency of GPR183⁺ cells to adult TEC^Wt and TEC^ΔSox4 thymic extracts (n=3 mice/condition). 100nM of 7α,25-HC was used as the positive control.

(H) Thymic expression analysis of indicated genes of TEC^WT and TEC^ΔSox4 mice by RT-qPCR normalized to mTEC1 specific gene Ccl21a (n=12 mice per genotype). P values determined by unpaired t-test.

Data in (A-H) are representative of at least 3 independent experiments. P values determined by unpaired t-test (all, except H) or by two-way ANOVA (H).

To assess epithelial Sox4 function on Tγδ17 development, we generated Foxn1^Cre;Sox4^fl/fl (TEC^ΔSox4) mice that lack Sox4 in epithelial cells in the thymus (Figure S4A). Gross thymic architecture was maintained and αβ thymocytes developed normally in TEC^ΔSox4 mice (Figure 4B, S4B and S4C). In sharp contrast, there were reductions in the numbers and proportions of Vγ2⁺ Tγδ17 cells in the thymus, sLN and skin (Figures 4D, 4C and 4E). The thymic alterations in Tγδ17 cells were age-dependent, more pronounced in neonates. Vγ2⁺ and Vγ4⁺ cells were differently affected (Figure 4C), with only the maturation of Vγ2⁺ Tγδ17 cells impaired. In adults (>8 wks old), defects in Vγ2⁺ Tγδ17 cells persisted in peripheral tissues including skin (Figures 4D and 4E), however accumulation of CCR6⁺ Vγ2⁺ T cells in the thymus of TEC^ΔSox4 normalized (Figure S4D). To determine whether oxysterol production was perturbed in TEC^ΔSox4 mice, we first performed the GPR183-mediated transwell migration assay with test cells that were transduced with Gpr183 expression vectors³⁰ (Figure 4F) and thymic extracts from WT and TEC^ΔSox4 mice. Direct quantitation of oxysterols in mouse tissues is not technically feasible due to the amount of tissue required.¹⁷ The in vitro migration assay is extensively used as specific and sensitive approach to measure GPR183 ligands.³⁰ This bioassay showed a reduced migration of test cells to extracts from TEC^ΔSox4 mice and indicated that they contained diminished amounts of GPR183 ligands (Figure 4G). Expression analysis of Ch25h and Cyp7b1 showed that thymus from TEC^ΔSox4 mice had less transcripts (Figures 4H and S4E). In comparison, expression of the chemokines Ccl21a and Ccl20 were only marginally altered with the loss of Sox4 in TECs, further confirmed using immunofluorescence assays (Figures S4E and S4F). Together, these results demonstrate that epithelial Sox4 controls Vγ2⁺ Tγδ17 cell maturation, in part by impacting Ch25h and Cyp7b1 expression in mTEC1.

Keratinocytes produce oxysterols

Equipping Tγδ17 thymocytes with GPR183 for export to skin predicts a corollary skin cellular network centered on oxysterol production that is critical for dermal Vγ2⁺ Tγδ17 cell function. We first determined oxysterol production in the skin using the transwell migration assay as above, comparing WT and Ch25h-deficient skin cell extracts. GPR183 transduced lymphocytes migrated to skin extracts in a concentration dependent manner, and the GPR183-dependent migration was not observed with skin extracts from Ch25h^−/− mice (Figure 5A). These results demonstrate that skin cells produce functional 7α,25-HC.

Figure 5. Epidermal-derived oxysterols maintain skin Tγδ17 cells.

(A) Relative migration of GPR183⁺ cells in a transwell assay to ear extracts from adult WT and Ch25h^−/− mice (n=3 mice per genotype). 100 nM of 7α,25-HC was used as the positive control.

(B) Analysis of Ch25h expression by RT-qPCR of dermal and epidermal cells isolated from adult WT ear skin post topical Imiquimod application for two days (n≥3 per time point).

(C) Relative migration of GPR183⁺ M12 cells in a transwell assay to ear extracts from adult WT mice treated for 1d with IMQ (n≥8 mice per genotype). 100 nM of 7α,25-HC was used as the positive control.

(D) Representative IF images (from 5 experiments) of ear skin from Ch25h reporter mice stained with DAPI, anti-RFP (left, to detect tdTom reporter) and anti-TCRδ (blue, right) Abs. HF, hair follicle.

(E) Representative flow cytometric profiles of γδ T cell subsets from adult Tcrd^Cre/ERT2 (iTCRδ^Wt) and Tcrd^Cre/ERT2;Gpr183^fx/fx (iTCRδ^ΔGpr183) ear skin (top). Summary of the compiled relative frequencies of dermal Vγ2⁺ and Vγ4⁺ (Vγ1.1⁻Vγ2⁻) T cells in induced Gpr183-deficient mice, compared to WT, set as 1 (bottom) (n≥ mice per genotype).

(F) Representative flow cytometric profiles of γδ T cell subsets from adult iTCRδ^Wt and iTCRδ^ΔGpr183 cervical LNs (top). Summary of relative frequencies of dermal Vγ2⁺ and Vγ4⁺ T cells in induced Gpr183-deficient mice, compared to WT, set as 1 (bottom). n≥7 mice per genotype.

Data in (A-F) are representative of at least 3 independent experiments. P values determined by unpaired t-test.

Ch25h expression is upregulated upon TLR signaling.³⁵ To identify skin cells that produce oxysterols we first determined the stromal cell fraction that inducibly transcribes Ch25h. In myeloid cells, Ch25h expression is rapidly upregulated with minimal expression at resting states.⁵ While basal amounts of Ch25h transcripts were detected in both the dermal and epidermal cell fractions, cutaneous treatment with the topical inflammatory TLR7/8 ligand Imiquimod (IMQ) upregulated Ch25h expression preferentially in epidermal keratinocytes (Figure 5B) and GPR183 ligand in the skin (Figure 5C). Immunofluorescence imaging of Ch25h-reporter mice identified basal keratinocytes⁹ termed interfollicular epidermal (IFE) keratinocytes and hair follicle (HF) outer bulge located at the dermal-epidermal border as the likely oxysterol depot in resting mice (Figure 5D), consistent with published skin scRNAseq datasets.^36,37 Dermal γδ T cells primarily localized with HF that are sheathed with IFE keratinocytes at the upper dermal-epidermal interface (Figures 5D, S5A and S5B), although other positioning is documented.³⁸

Vγ2⁺ Tγδ17 cells are skin resident, and once established after birth they can be maintained without inputs from peripheral lymphoid tissues. To determine whether continuous sensing of oxysterols by GPR183 is required for dermal Tγδ17 cell maintenance we acutely depleted Gpr183 from γδ T cells by administration of Tamoxifen (Tam) to Tcrd^Cre/ERT2;Gpr183^fl/fl mice (iTCRδ^ΔGpr183). We confirmed that the Tam-mediated recombination was efficient (over 70% of skin γδ T cells had CRE activity) and specific (no CRE activity detected in skin αβ T cells, Figure S5C). Within 6 days of Tam-treatment there was a significant loss of dermal Vγ2⁺, but not Vγ4⁺, Tγδ17 cells (Figure 5E). This effect was muted in sLNs (Figure 5F). Correspondingly, acute removal of Ch25h from keratinocytes (by Tamoxifen treatment of Krt5^Cre/ERT2;Ch25h^fl/fl mice³⁹ also specifically diminished the frequency of Vγ2⁺ Tγδ17 cells (Figure S5D). Together, these results show that basal keratinocytes express Ch25h that can be modulated by inflammatory cues, and that the GPR183-oxysterol ligand axis is necessary for optimal maintenance of skin resident Vγ2⁺ Tγδ17 cells.

Diet-derived oxysterol determines severity of Vγ2⁺ Tγδ17 cell-mediated psoriasis

We and others have shown that Vγ2⁺ (but not Vγ4⁺) Tγδ17 cells are absolutely required for psoriasis driven by cutaneous treatment with IMQ.^10,40 Increased dietary intake of fat and lipid can sensitize mice to develop more severe IMQ-mediated psoriasis,⁴¹ correlating with enhanced activities of Tγδ17 cells and providing one plausible explanation for a correlation between obesity and severity of tissue inflammatory disorders in humans, including psoriasis and atopic dermatitis.⁴² Molecular pathways controlling diet regulated T3L function in mucocutaneous tissues are unknown.

To test whether dietary cholesterol calibrates proximal Tγδ17 inflammatory responses via GPR183, we fed mice with chow supplemented only with cholesterol (2%, High Cholesterol Food, HCF) for 2 wks. GPR183 ligand and its precursor were increased in the skin of mice fed HCF as assessed by the migration assay (Figures 6A and S6A), correlating with increased expressions of Ch25h and Cyp7b1 in skin cells (Figure 6B). Prolonged exposure (8 wks) to 2% HCF led to ear thickening (Figure S6B) and increased basal in vivo IL-17 production by dermal Vγ2⁺ Tγδ17 (Figure S6C). Psoriatic response to IMQ topical administration was exacerbated in mice fed 2% HCF for 2 wks, as reflected by the two standard measures of disease severity in this model, increased ear thickness and epidermal hyperplasia resulting from overt inflammatory response (Figures 6C and 6D). We took advantage of the normal numbers of Vγ2⁺ Tγδ17 cells in Gpr183^−/− mice (Figure 2B, contrasted to DKO mice and in the Gpr183 acute depletion model), to assess whether GPR183-oxysterol sensing is required for the psoriatic response controlled by Tγδ17 cells. Ch25h-deficient mice are not an ideal model for this purpose as CH25H has pleiotropic effects on inflammatory responses in most tissues.^5,43 Mice lacking GPR183 only marginally responded to IMQ treatment (Figure 6E) and HCF did not enhance the response (Figure 6F), suggesting that sensing of 7α,25-HC is required for Tγδ17 tissue effector function. Surprisingly, at the end of IMQ treatment in the HCF cohort, there was a significant depletion of Vγ2⁺ T cells in the skin and LNs of IMQ-treated Gpr183^−/− mice, with attendant decreases in Vγ2⁺ Tγδ17 cells in these tissues (Figures 6G and 6H). The fetal-origin Vγ4⁺ Tγδ17 cells were also decreased in the skin of IMQ-treated Gpr183^−/− mice, but no differences were observed in sLNs. DETCs were unaltered and comparable to untreated WT mice (Figure S6D). Cholesterol-enhanced psoriasis required GPR183 expression on T cells and dependent on 7α,25-HC production, since mice lacking GPR183 in T cells (Rorc^CreGpr183^flox/flox), and mice deficient for CH25H showed blunted skin inflammation upon 2% HCF and IMQ topical administration (Figures S6E and S6F). These results indicate that Gpr183 is required to maintain tissue Tγδ17 cells in sustained inflammatory settings, particularly with high concentrations of tissue cholesterol metabolites.

Figure 6. Dietary cholesterol exacerbates Imiquimod-induced psoriasis that is dependent on GPR183 oxysterol cognition.

(A) Relative migration of GPR183⁺ cells to ear extracts from mice on NF or 2% HCF diets for 2 wks (n≥4 per diet group), measured in the transwell migration assay.

(B) Analysis of Ch25h and Cyp7b1 expression by RT-qPCR of ear skin epidermal cells of adult WT mice on different diets (NF, normal food or 2% HCF, high cholesterol food) for 2–4 weeks (n=3 per diet group). 100 nM of 7α,25-HC was used as the positive control.

(C) Relative ear-skin thickness (set at 1 at d0) measurements for 6 days from of IMQ-treated WT mice on NF or 2% HCF diet (n≥11 per diet group).

(D) Representative hematoxylin-and-eosin staining of ear skin from WT mice treated with IMQ for 5d as in (c) (n=5 per diet group). D, dermis; E, epidermis. Scale bar is 100 μm.

(E-F) Relative ear-skin thickness measurements of IMQ-treated WT and Gpr183^−/− mice on NF diet (n≥6 mice per genotype) (E) or on 2% HCF diet (n=5 mice per genotype) (F).

(G-H) Representative flow cytometric profiles of intracellular IL-17 (icIL-17) amounts in Vγ2⁺ T cells from WT and Gpr183^−/− skin (G) and cervical LNs (H) of mice (2% HCF diet) treated daily for 5 d with IMQ (top rows). Summary of frequencies of icIL-17⁺ cells expressing Vγ2⁺TCR or Vγ4⁺TCR (bottom rows) amongst hematopoietic cells (CD45⁺) determined in the same studies (n=5 mice per genotype), at the end of IMQ treatment.

Data in (A-H) are representative of at least 3 independent experiments. P values determined by unpaired t-test (A, G, H) or by two-way ANOVA (B, D-F).

Together, these findings demonstrate that one consequence of a diet rich in cholesterol is increased susceptibility to overt psoriasis. Changes in dietary cholesterol metabolite composition in the skin are sensed by GPR183 on dermal T cells, calibrating their pathogenic effector activities.

Discussion

Based on the long-standing link between obesity and heightened inflammatory conditions in multiple human diseases, observational data have been interpreted to provide correlation between types of diet and lymphocyte-mediated inflammatory diseases. However, high fat/cholesterol-regulated lymphoid molecules critical for immune responses are unknown, aside from the secreted, generic inflammatory cues (e.g. TNF, IL-6) that affect nearly all cells in the body. Moreover, investigation into the lipid metabolism-T cell axis has been primarily focused on intracellular synthesis of cholesterol and its metabolite for both lipid anabolic program and antigen-receptor signaling.^44–50 Tγδ17 thymocytes (both Vγ4⁺ and Vγ2⁺) have been reported to accumulate intracellular lipid droplets and upon exit to periphery they appear dependent on cholesterol as the main energy source.⁵¹ While extracellular lipoprotein modulation of T cell function in vitro and in vivo has been documented,^52–58 mechanistic links amongst diet-derived (i.e. not endogenously synthetized) metabolites, specific sensors on lymphocytes, and direct immune functional outcomes have been lacking. Here we show that high cholesterol from the diet is associated with increased production of oxysterols and that in turn worsens DAMPelicited, Vγ2TCR⁺ Tγδ17 cell-driven psoriatic responses. Remarkably, this oxysterol sensing capacity is “learned” during thymic differentiation, as cells incapable of sensing the ligands for GPR183 and CCR6 do not mature. The gene network of oxysterol production in the thymus is conserved in the human thymus³⁴, suggesting that human GPR183⁺ T cells will exhibit shared function with the mouse counterparts.

Our data for the first time reveal the possibility of dedicated mTECs that regulate innate T3L, paralleling the regulatory circuit required for innate Type 2 cytokine producers.^31,33 The developmentally programmed GPCR function described here is unique and requires a specific anatomical thymic niche composed of Sox4⁺ mTECs that are a subset of mTEC1. We have previously shown that Sox4 regulates expression of T3L signature genes and is required for Tγδ17 development in a thymocyte-intrinsic manner.²¹ That Sox4 is also required in the epithelial cells to ensure maturation of Tγδ17 thymocytes in trans provides a pleasing symmetry of coordinated thymocyte-stromal interactions. Given that Sox4 also regulates other thymic-derived, αβTCR⁺ innate T3L development, such as iNKT17 and MAIT17,²¹ it would be of interest to determine whether their maturation would also be controlled by Sox4+ mTECs. Together, the findings reveal a highly organized mTEC subset functionality geared for effector type-specific, non-conventional thymocyte maturation. Understanding how distinct TF networks impart mTEC functionality is central to redefining T cell education processes in the thymus, beyond the conventional αβ T cell selection process.

How mobile innate lymphoid sentinels are positioned in barrier tissues is mostly unknown, especially in a homeostatic state with no overt inflammation. The dermis is sparsely populated with immune cells and the density of dermal commensal bacteria is very low.⁵⁹ An efficient surveillance system would need specialized cells with a guidance system that seek out anatomical hubs concentrated with molecular cues indicative of tissue dysbiosis. GPCRs mediate immune cell positioning in tissues. Yet, except in relatively rare cases,^60,61 single GPCR deficiency does not translate to obvious cell positioning defects, as other chemotactic cues can guide cell delivery to sub-anatomical niches, and random walk eventually allows immune cells to reach their intended localization even without sensing migratory gradient. Hence, competitive experimental settings (such as adoptive transfer and mixed bone marrow chimera) are needed to reveal the requirements for in vivo localization^62–65 as previously shown, among others, for CCR6 and GPR183.^14,66 Acute alterations of GPR183-oxysterol interaction in skin impact Vγ2⁺ Tγδ17 cell accumulation in the dermis, likely due to immediate cell-to-cell competition for anatomically segregated ligand depots in the tissue. Whether similar processes are operational in human skin are unknown. In adults, the majority of IL-17 producing innate-like T cells appears to be MAIT.⁶⁷ GPR183 is detected on some human skin MAITs and the GPR183 ligand is produced by the human skin.⁶⁸ MAIT17 are thus the best functional correlates of mouse cutaneous Tγδ17 cells. While Tγδ17 cells have been reported to be rare in adult human mucocutaneous tissues^18,68 fetal Tγδ17 cells develop in the late first trimester of pregnancy.⁶⁹ Whether this early wave of Tγδ17 cells populate infant skin that then decay over time is an active area of investigation.

In addition to the conserved oxysterol gene network in human thymus, human genetic variations in Gpr183 have been linked to psoriasis,⁷⁰ inflammatory bowel diseases (IBD)⁷¹ and Type 1 diabetes.⁷² A single nucleotide polymorphism in Gpr183 increases the risk for IBD⁷¹ and in mouse, Gpr183 deficiency impairs intestinal ILC3 localization.¹¹ Moreover, an inability to sense GPR183 ligand protects from a murine model of colitis, suggesting that mechanisms underpinning tissue damage driven by heightened GPR183-mediated migration and signaling is conserved across mucocutaneous tissue barriers. Ch25h expression and 25-HC production are controlled by TLR ligands in a Type I IFN–dependent fashion,⁷³ thereby modulating GPR183 ligand concentration in tissues and potentially serving as a rheostat of Tγδ17 cell (and likely other T3L) activation. Whether inflammatory cascades also impact homeostatic GPR183 ligand concentration is largely unknown. While commensal bacteria in mucocutaneous tissues can influence a wide range of immune processes, they do not seem to modulate Ch25h induction or Vγ2⁺ Tγδ17 cell maintenance.^9,11 Rather, our data support the possibility that dietary metabolites, and other life essential derivatives, calibrate tissue GPR183⁺ T3L function.

STAR METHODS

RESOURCES AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Andrea Reboldi (andrea.reboldi@umassmed.edu).

Materials availability

Ch25h^tdTom mouse line generated in this study is available from the lead contact with a completed materials transfer agreement.

Data and code availability

Single-cell RNAseq data γδ thymocytes have been deposited at GEO and are publicly available as of the date of publication. Accession numbers are listed in the key resources table. This paper analyzes existing, publicly available single-cell RNAseq data of TEC subsets. The accession number for this dataset is listed in the key resources table. Microscopy data reported in this paper will be shared by the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. This paper does not report original code.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Syrian Hamster anti-mouse CD3e Biotin	eBioscience	Cat#13003382 RRID:AB_1470754
Rat anti-mouse CD8α Biotin	Biolegend	Cat#100704 RRID:AB_312743
Rat anti-mouse CD11b Biotin	Biolegend	Cat#101204 RRID:AB_312787
Armenian Hamster anti-mouse CD11c Biotin	eBioscience	Cat#13011485 RRID:AB_466363
Rat anti-mouse CD19 Biotin	BD Biosciences	Cat#553784 RRID:AB_395048
Rat anti-mouse Gr-1 Biotin	Biolegend	Cat#108404 RRID:AB_313369
Mouse anti-mouse NK1.1 Biotin	Biolegend	Cat#108704 RRID:AB_313391
Rat anti-mouse Ter-119 Biotin	BD Biosciences	Cat#553672 RRID:AB_394985
Armenian Hamster anti-mouse TCRβ Biotin	eBioscience	Cat#13596185 RRID:AB_466820
Armenian Hamster anti-mouse TCRγ/δ Biotin	Biolegend	Cat#118103 RRID:AB_313827
Rat anti-mouse CD4 Biotin	eBioscience	Cat#13004385 RRID:AB_466334
Armenian Hamster anti-mouse TCRβ BV605	Biolegend	Cat#109241 RRID:AB_2629563
Armenian Hamster anti-mouse TCRγ/δ PerCP-eFluor 710	eBioscience	Cat#46571182 RRID:AB_2016707
Armenian Hamster anti-mouse Vγ2 PE-Cy7	eBioscience	Cat#25582882 RRID:AB_2573474
Armenian Hamster anti-mouse Vγ1.1 BV421	BD Biosciences	Cat#566308 RRID:AB_2739676
Rat anti-mouse CD24 BV605	Biolegend	Cat#101827 RRID:AB_2563464
Rat anti-mouse CCR6 Alexa Fluor 647	BD Biosciences	Cat#557976 RRID:AB_2228793
Armenian Hamster anti-mouse CD27 PE	eBioscience	Cat#12027183 RRID:AB_465615
Rat anti-mouse CD4 BV650	Biolegend	Cat#100546 RRID:AB_2562098
Rat anti-mouse CD44 BV711	Biolegend	Cat#103057 RRID:AB_2564214
Rat anti-mouse IL-17A PerCPCy5.5	eBioscience	Cat#45717782 RRID:AB_925753
Rat anti-mouse IFNγ PE	Biolegend	Cat#505808 RRID:AB_315402
Rat anti-mouse CD8β APC/Cyanine7	Biolegend	Cat#126620 RRID:AB_2563951
Rat anti-mouse CD45 BV785	Biolegend	Cat#103149 RRID:AB_2564590
Rat anti-mouse CD49f FITC	Biolegend	Cat#313606 RRID:AB_345300
Rat anti-mouse Sca-1 BV605	Biolegend	Cat#108134 RRID:AB_2650926
Rat anti-mouse CD71 PE-Cy7	Biolegend	Cat#113811 RRID:AB_2203383
Rat anti-mouse CD326 PE	Biolegend	Cat#118206 RRID:AB_1134172
Rat anti-mouse CD326 PE-Cy7	Biolegend	Cat#118216 RRID:AB_1236471
Rat anti-mouse I-A/I-E APC	Biolegend	Cat#107614 RRID:AB_313329
Armenian Hamster anti-mouse CD80 PerCPCy5.5	Biolegend	Cat#104721 RRID:AB_893406
Rat anti-mouse Ly51 FITC	Biolegend	Cat#108305 RRID:AB_313362
Rat anti-mouse CD104 PE	Biolegend	Cat#123610 RRID:AB_2563544
UEA-1 biotin	Vector Labs	Cat#B1065 RRID:AB_2336766
Rat anti-mouse Scart2	J. Kisielow	Cloned hybridoma 25A2
Goat anti-Rat IgG PE	Southern Biotech	Cat#305009 RRID:AB_2795834
Rat anti-mouse Anti-CD16/32	Biolegend	Cat#101330 RRID:AB_2561482
Rabbit anti-GFP	Invitrogen	Cat#A11122
Rabbit anti-RFP	Rockland	Cat#600401379 RRID:AB_11182807
Rat anti-mouse SCART2 Alexa Fluor 647	Muzaki et al.74	Cloned hybridoma 41G1
Rat anti-mouse AIRE Alexa Fluor 488	eBioscience	Cat#53593480 RRID:AB_10852560
Cy3-conjugated donkey anti-rabbit	Jackson ImmunoResearch Laboratories	Cat#711165152 RRID:AB_2307443
Streptavidin Alexa Fluor 450	eBioscience	Cat#48431782 RRID:AB_10359737
Chemicals, peptides, and recombinant proteins
Fetal Bovine Serum	Sigma-Aldrich	Cat#F0926
Phorbol 12-myristate 13-acetate (PMA)	LC Laboratories	Cat#P-1680
Ionomycin Calcium Salt	Sigma-Aldrich	Cat#I3909
GolgiStop (Monensin)	BD Biosciences	Cat#512301KZ
GolgiPlug (Brefeldin A)	BD Biosciences	Cat#512092KZ
DNase I, grade II	Roche	Cat#10104159001
Liberase TL	Roche	Cat#05401020001
Liberase TM	Roche	Cat#05401119001
Hylauronidase	Sigma-Aldrich	Cat#H3506
Trizol	Ambion Life Technologies	Cat#15596018
Tamoxifen	Sigma-Aldrich	Cat#T5648
Peanut Oil	Sigma-Aldrich	Cat#P2144
Hepes (1 M)	Gibco	Cat#15630080
0.25% Trypsin in HBSS w/o Calcium and Magnesium	Corning	Cat#25050CI
EDTA 500mM, pH 8.0	Teknova	Cat#E036
RLT buffer (RNeasy Lysis Buffer)	Qiagen	Cat#79216
2-mercaptoethanol	Sigma-Aldrich	Cat#6250
Percoll	GE Healthcare	Cat#17089101
DAPI	Sigma Aldrich	Cat#D9542
7α,25-dihydroxycholesterol	Avanti Polar Lipids	Cat#700080P
25-hydroxycholesterol	Avanti Polar Lipids	Cat#700019P
Imiquimod Cream 5%	Perrigo	Cat#45802036862 NDC#45802036862
Critical commercial assays
Cytofix/Cytoperm Fixation/Permeabilization Kit	BD Biosciences	Cat#554714
Fixable Viability Dye eFluor 780	eBioscience	Cat#65086514
Fixable Viability Dye eFluor 506	eBioscience	Cat#65086618
RNeasy Micro kit	Qiagen	Cat#74004
RNeasy Mini kit	Qiagen	Cat#74104
SuperScript First-Strand Synthesis System for RT-PCR	Invitrogen	Cat#11904018
iQ SYBR green Supermix	Bio-Rad	Cat#1725270
Fluoromount-G	SouthernBiotech	Cat#0100–01
Deposited data
Single-cell RNA Sequencing of γδ thymocytes	This Paper	GEO: GSE221488
Single-cell RNA Sequencing of TEC subsets	Cowan et al.75	GEO: GSE131368
Experimental models: Cell Lines
M12-EBI2-GFP	Gift from J. Cyster	Kelly et al.30
Human Embryonic Kidney Cells (HEK293T)	ATCC	Cat#CRL-3216
Experimental models: Organisms/strains
Mouse: C57BL/6J	The Jackson Laboratory	000664
Mouse: Ccr6−/−	The Jackson Laboratory	005793
Mouse: Ch25h−/−	The Jackson Laboratory	016263
Mouse: TcrdCre/ERT2	The Jackson Laboratory	031679
Mouse: Foxn1Cre	The Jackson Laboratory	018448
Mouse: RorcCre	The Jackson Laboratory	022791
Mouse: Krt5Cre/ERT2	The Jackson Laboratory	029155
Mouse: Sox4Egfp	MMRRC	030033-UCD
Mouse: Gpr183+/Egfp	Pereira et al.14	N/A
Mouse: Gpr183fl/fl	Pereira et al.14	N/A
Mouse: Sox4fl/fl	Malhotra et al.21	N/A
Mouse: Ch25hfl/fl	Ceglia et al.39	N/A
Mouse: Ch25htdTom	This paper	N/A
Oligonucleotides
See Table S1 for list of quantitative RT-PCR primers	N/A	N/A
Recombinant DNA
MSCV-IRES-Thy1.1	Addgene	ID: 17442
Software and algorithims
FlowJo v10.8 software	Tree Star	https://www.flowjo.com/solutions/flowjo/downloads
FACSDiva v7.0 software	BD	https://www.bdbiosciences.com/en-us/products/software/instrument-software/bdfacsdiva-software
ZEN 3.1	Carl Zeiss Microscopy	https://www.zeiss.com/microscopy/en/products/software/zeiss-zen.html
Adobe Illustrator 2022	Adobe Systems	N/A
ImageJ	NIH	https://imagej.nih.gov/ij/
Prism 9	GraphPad Software	https://www.graphpad.com/scientificsoftware/prism/
Seurat R v2.3	Butler et al.76	https://satijalab.org/seurat/
Seurat R v3.2	Stuart et al.77	https://satijalab.org/seurat/
Other
Standard chow diet (Prolab IsoPro RMH 3000)	LabDiet	5P76
2% high cholesterol diet	Envigo	TD.200179

EXPERIMENTAL MODELS AND SUBJECT DETAILS

Mice

Ccr6^−/− (Stock no: 005793), Ch25h^−/− (Stock no: 016263), Tcrd^Cre/ERT2 mice (Stock no: 031679), Foxn1^Cre mice (Stock no: 018448), C57BL/6J mice (Stock no: 000664), Rorc^Cre mice (Stock no: 022791) and Krt5^Cre/ERT2 mice (Stock no: 029155) were from Jax Laboratories. Gpr183^+/Egfp (reporter insertion creates a null allele) and Gpr183^fl/fl mice were described previously^14,78, Sox4^Egfp mice were sourced from MMRRC (Stock no: 030033-UCD), Sox4^fl/fl mice were described previously²¹, originally from V. Lefebvre⁷⁹ and Ch25h^fl/fl were described previously.³⁹ Ch25h^tdTom knock-in animals were generated using crRNA (gtgggacatagtctcagcat) and tracrRNA ordered from IDT, which target around the start codon of Ch25h. To make the donor DNA, a 5’ homology arm (0.8kb) and 3’ homology arm (0.8kb) were generated from mouse genomic DNA by PCR and cloned into vector pCR-Blunt II-TOPO to flank the tdTomato-polyA insert. RNP assembly followed standard protocol of IDT. sgRNA, circular donor DNA, and Cas9 protein were provided to the Gladstone Gene Targeting Core for C57BL/6J blastocyst microinjection. 2/14 mice were confirmed by sequencing to have tdTomato-polyA inserted immediately after the start codon of Ch25h. These two founders were bred to C57BL/6J mice and germline transmission was confirmed by allele-specific PCR. Males and females were used for experiments, but sex-matched within an experiment. No differences were observed between sexes. Ages of mice used for experiments are indicated in Figure Legends. Animals were randomly allocated to experimental groups. All mouse procedures were approved by the University of Massachusetts Medical School IACUC.

METHOD DETAILS

Diets

Mice were either fed a standard chow diet (Prolab IsoPro RMH 3000 5P76) or a high cholesterol diet (2% cholesterol added to the Prolab RMH 3000 5P76; Envigo TD. 200179, customized diet) for the duration of the experiment.

Tamoxifen treatment

Tamoxifen (Sigma) was dissolved in Peanut Oil (Sigma) oil at 25 mg/mL and 5mg of tamoxifen (250 μg/body weight) was orally gavaged 3 times into mice (6 to 8 weeks old) every other day.

Cell isolation

Skin single cell suspensions were prepared from mouse ears. Ears were depilated using Nair application for two minutes, followed by gentle removal of cream and rinsing of tissues using PBS, as previously described.⁹ For total skin cell isolation, as previously described⁹, depilated mouse ears were peeled into dorsal and ventral halves, chopped finely using scissors, and digested with 1 U/mL Liberase TL (Roche) + 0.5 mg/mL Hylauronidase (Sigma-Aldrich) + 0.05 mg/mL DNAse (Roche) dissolved in HBSS (Corning) + 5% FCS (Sigma-Aldrich) + 1 mM Hepes (Gibco) on a stir plate for 90 minutes at 37°C. To stop digestion, 10 mM EDTA (Teknova) was added. Cells were filtered through 100 μm cell strainer, rinsed with HBSS (Corning) + 5% FCS (Sigma-Aldrich) + 1 mM Hepes (Gibco) + 1% Penicillin/Streptomycin (Gibco), spun down, filtered through 70 μm cell strainer, rinsed again, spun down, and plated for antibody staining for flow cytometry or for cytokine stimulation. For dermal and epidermal cell isolation, depilated ear tissues were peeled into dorsal and ventral halves and each placed dermal side down to float on 0.25% Trypsin (Corning) with 1 mM EDTA (Teknova) for 1 hr at 37°C, as previously described.⁸⁰ Following incubation, epidermis was gently scraped off of dermal layer of ear halves using forceps, filtered through 100μM cell strainer into RPM1-1640 (Gibco) + 10% FCS (Sigma-Aldrich) + 1 mM Hepes (Gibco) + 1% Penicillin/Streptomycin (Gibco), and spun down for subsequent procedures. The dermal layer was collected and homogenized in RLT buffer (Qiagen) + 1% 2-mercaptoethanol (Sigma-Aldrich) for downstream analysis. For thymic epithelial cell isolation, single-cell thymic suspensions were prepared, and TEC isolated for cell sorting as previously described.⁷⁵ In brief, the thymus was cut in small pieces and digested with 63μg/mL Liberase TM (Roche) + 20 μg/ml DNaseI (Roche) while shacking for 40 min at 37°C. Cells were then resuspended in 1.115 g/mL isotonic Percoll (GE Healthcare), overlaid with 2 mL of 1.065 g/mL isotonic Percoll and 2 mL of PBS. Samples were centrifuged at 2700 RPM for 20 min at 4°C with the brake off. The thymic epithelial cells were enriched at the interface between the Percoll and the PBS layers. The isolated cells were then stained for cell sorting as described below.

Antibodies and Flow Cytometry

Flow cytometry staining was performed in 96-well microtiter plates. Antibody cocktails were diluted in PBS (GIBCO) + 2% FBS (Sigma-Aldrich) + 2 mM EDTA (Teknova), and cells were stained in 50 μL for 20 min on ice. With the exception of SCART2 antibody (25A2 hybridoma) kindly donated by the Kisielow group, all of the following antibodies were purchased from BD Biosciences, Biolegend or eBioscience: CD3ε (500A2), CD4 (RM4–5), CD8α (53–6.7), CD8β (YTS156.7.7), TCRβ (H57–597), TCRδ (GL3), Vγ2 (UC3–10A6), Vγ1.1 (2.11), CD27 (LG.3A10), CCR6 (140706), CD24 (M1/69), CD44 (IM7), CD45 (30-F11), IFNγ (XMG1.2), IL-17A (7B7), CD34 (RAM34), CD49F (GoH3), Sca1 (D7), CD326 (EpCAM) (G8.8), I-A/I-E (M5/114.15.2), CD80 (16–10A1) and Ly51 (6C3). Biotinylated UEA-1 (B-1065) was from Vector Labs. Anti-rat IgG (Southern Biotechnology Associates) was used as a secondary reagent for SCART2 detection. All samples were labeled with Fixable Viability Dye (Thermofisher) to exclude deal cells from all analysis. For intracellular cytokine staining, cells were stimulated in vitro with 10 ng/mL phorbol myristate acetate (PMA) + 1 μM Ionomycin (both Sigma-Aldrich) in the presence of GolgiPlug and GolgiStop (BD Biosciences) for 4 hours at 37°C, surface stained, LIVE/DEAD labeled, fixed/permeabilized with Cytofix/Cytoperm buffer (BD Biosciences) and then stained for indicated intracellular cytokines. Data were acquired on a BD LSRII cytometer or FACSAria (BD Biosciences) and analyzed using FlowJo (Treestar).

Cell sorting

For γδ thymocyte sorting, thymocytes from 12-day old mice were stained and sorted as TCRδ⁺/CD3⁺/TCRβ⁻/CD4⁻/CD8⁻ with a 70μm nozzle. For TECs cell sorting, single cells thymic suspensions from 4 weeks old mice were stained for CD45, EpCAM, Ly51, CD80, I-A/I-E, UAE-1 and sorted as previously described.⁷⁵ All cells were sorted on a BD FACSAria II.

Histology and immunofluorescence

For H&E staining, muzzle tissue was first fixed in 10% neutral-buffered formalin for 24 hr, and then paraffin embedded, sectioned, and stained by the UMMS DERC Morphology Core. For immunofluorescence, tissues were fixed in 4% paraformaldehyde (J.T. Baker), 0.53 M L-Lysine, 2.1 mg/ml sodium m-periodate, 0.024 N NaOH in PB for at least 2 hours at 4°C, washed three times for 10 min in PB, then moved to 30% sucrose in PBS overnight. Tissues were flash frozen in Tissue-Tek Cryomold (VWR) the next day, and 7-mm sections were cut and then dried for 1 hour before staining. Sections were rehydrated in PBS, blocked, and then stained in primary antibody overnight at 4°C and stained for subsequent steps for 2 hours at room temperature. For thymus, slides were blocked with 5% normal mouse serum + 5% normal rat serum in PBS, 0.3% Triton-X100 (Sigma), 0.2% Bovine Serum Albumin (BSA) and 0.1% sodium azide. All the antibodies were diluted in the same buffer containing 2% normal mouse serum + 2% normal rat serum. For skin, slides were permeabilize for 10 minutes with 0.1% Triton X100 (Sigma) in PBS, and then blocked for 1 hr in blocking buffer (2.5% normal donkey serum, 2.5% normal goat serum, 2.5% normal mouse serum, 2.5% normal rat serum, 1% BSA, 1% Gelatin, 0.3% Triton X-100). All the antibodies were diluted in the same buffer. Sections were stained with the following primary antibodies: Rabbit anti-GFP (polyclonal, Life Technologies), Rabbit anti-RFP (polyclonal, Rockland), Alexa647-conjugated anti-SCART2⁷⁴, Biotinylated anti-TCRδ (GL3, eBioscience), Alexa488-conjugated anti-AIRE (5H12, eBioscience). Sections were then stained with the following secondary antibodies: Cy3-conjugated donkey anti-rabbit (Jackson Immunoresearch), Alexa488-conjugated streptavidin (Invitrogen) and DAPI (BioLegend). Images were obtained with a Zeiss AxioObserver.Z1 (Carl Zeiss) inverted microscope and were analyzed by imaging processing software ImageJ (NIH).

Imiquimod model of psoriasis

Mice between 6 and 12 weeks of age were anesthetized with isoflurane before treating each ear with 5 mg of 5% imiquimod cream (Imiquimod Cream 5%; Perrigo) daily for up to 5 days. Peripheral and central ear thicknesses were measured daily with a digital caliper (Mitutoyo).

Oxysterol chemoattractant bioassay

Mouse tissue extracts were prepared as previously described.³⁰ Briefly, organs were weighed and mashed in 10 times the volume in mg (100 mg/ml) of sterile chemotaxis media (RPMI + 0.5% fatty acid-free BSA) through a 70-mm filter. Clean supernatants were collected after centrifugation and tested for bioactivity by Transwell chemotaxis assays of an EBI2⁺ M12 B cell line.³⁰ 50 nM 7α,25-HC (Avanti Polar Lipids) was used as positive control. Relative amounts of in situ 25-HC was determined as described⁸¹. Briefly, skin extracts were incubated sequentially with HEK293T cells transfected with a MSCV retroviral construct encoding mouse Hsd3b7 (to remove in situ 7α,25-HC) and then with Cyp7b1 transfected HEK293T cells (to convert in situ 25-HC to 7a,25-HC). Supernatants from the latter were used in the M12 bioassay.

RNA extraction and Real-Time RT-PCR

For RT-qPCR analysis of the thymus, the organ was homogenized in RLT buffer (Qiagen) + 1% 2-mercaptoethanol (Sigma-Aldrich) using an Omni Tissue homogenizer. For RT-qPCR analysis of the skin epidermis and dermis, cells were isolated as described above. Epidermal and dermal cells were homogenized in RLT buffer (Qiagen) + 1% 2-mercaptoethanol (Sigma-Aldrich). RNA was isolated using the RNeasy Mini or Micro kit (Qiagen). RNA was converted to cDNA using oligo dT priming and SuperScriptIII (Invitrogen). qPCR was performed using iQ SYBR green Supermix and a CFX96 thermal cycler (Bio-rad), followed by thermal melt curve analysis to confirm specific amplification. Primers used in this study were synthesized by Integrated DNA Technologies and are reported in Table S1.

Single-cell RNAseq data acquisition

γδ thymocyte data generation was performed using the standard operating protocol of the Immunological Genome Project Consortium (Immgen.org). Sorted TCRδ⁺ thymocytes were processed through 10X Chromium (10X Genomics) and cDNA libraries were prepared following the recommended protocol for the Chromium Single Cell 30 Reagent Kit (v2 Chemistry). Post-processing and quality control were performed using the 10X Cell Ranger pipelines (10X Genomics). Estimated number of cells was 3,749 with 1,894 median genes per cell sequenced with 37,384 mean reads per cell. Seurat R package (v2.3.4) was used for further analysis with default parameters applied. TEC populations were sequenced as described before.⁷⁵ For the analysis of TEC single-cell RNA-Seq data, Seurat R package (v3.2.2) was used with default parameters. Using Seurat, the contribution of cell-cycle to variability was controlled for by regressing out expression of 94 genes previously shown to be associated with cell cycle activity. We performed cluster analysis at a resolution that enabled the identification of the four mTEC subsets, based on expression of selected marker genes.³¹

QUANTIFICATION AND STATISTICAL ANALYSIS

Statistical analysis of scRNAseq data was performed using R. Summary data from FACS analyses were analyzed in GraphPad Prism software using statistical tests indicated in Figure Legends. The mean of all samples in a group is used to represent the central tendency of the dataset, and all error bars represent SEM of biological replicates. Sample size was not determined prior to experimentation. The exact significance values are stated in all graphs and the number of biological replicates (n) is stated in the Figure Legends. No randomization of experiments was conducted. Experimenters were not blinded during performance or analysis of the experiments.