Receptor Mincle promotes skin allergies and is capable of recognizing cholesterol sulfate

Mincle Is a PRR in the Skin That Is Highly Inducible in Response to Trauma and Mediates the Early Inflammatory Response.

We first assessed changes in the gene expression profiles of the receptors involved in innate responses to skin injuries in mice. We modeled excisional cutaneous wounds using a punch biopsy instrument. Wounds were made through the epidermis, dermis, and s.c. tissue layers, leaving the fascia intact. Skin tissue samples containing the wounds were removed 24 h and 4 d after injury, and intact skin was used as a control. RNA was extracted from samples, and a whole-genome analysis was performed using an Affymetrix microarray platform (Agilent Technologies). The results revealed that skin injury induced an increase in the expression of a number of PRR genes belonging to different families of innate immunity receptors, such as Toll-like receptors (TLRs), C-type lectin receptors (CLRs), NOD-like receptors (NLRs), and RIG-I–like receptors (RLRs). The fold-changes in the transcript levels of innate immune PRR genes in wounded skin relative to levels in intact skin are presented in Fig. 1A (the complete dataset has been deposited in the GEO database under accession no. GSE76144). Interestingly, one of the genes of the CLR family, Clec4e (Mincle), was extremely inducible after wounding in comparison with the changes in the other innate immune receptors. Another gene of the C-type lectin receptor family, Clec4d (MCL), which is located in the CLR cluster next to Clec4e and is considered to have emerged via a gene duplication of Clec4e (13), was the second most inducible gene at both time points among the innate immune system receptor genes.

To investigate changes in the expression of Clec4e (Mincle) upon skin injury at the protein level, we compared Mincle expression in damaged skin tissue at different stages of the wound repair process using Western blotting. In accordance with the gene array data, the expression of the Mincle protein increased quickly after injury and had decreased to baseline levels by 12 d after wounding (Fig. 1B). It is noteworthy that the rapid induction of Mincle expression was observed not only after the physical injury of the skin but also in response to the single topical application of the irritant and sensitizer DNFB onto the abdomen skin of mice (Fig. 1C). Immunohistochemical studies revealed a population of Mincle-expressing cells in the dermis that had mononuclear histiocyte-like morphological characteristics and that appeared in the damaged skin as an early response to wounding or DNFB application (Fig. 1D). Only a small number of weakly stained Mincle-expressing cells can be found in intact skin. Double immunofluorescent staining performed 2 h after DNFB application revealed that the Mincle-expressing cells do not express Langerin or CD11c in detectable quantities. However, these cells possess phenotype markers of plasmacytoid dendritic cells (pDC), such as BST2 (PDCA-1) and CD45R (B220) (Fig. 1E). Taken together, these findings indicate that Mincle is a receptor that is highly inducible in response to skin tissue damage.

To evaluate the role of Mincle in early inflammatory response, we measured concentration of proinflammatory mediators in skin at the early time point (12 h) after trauma using Mincle-deficient mice and wild-type mice. We determined that, after skin wounding or after a single topical application of DNFB, the early secretion of proinflammatory mediators in the absence of Mincle was significantly impaired (Fig. 1F). However, the absence of Mincle did not influence the dynamics of wound closure (Fig. S1).

The findings are consistent with those of recently reported studies that have shown increased Mincle expression in the brains of humans and rodents after brain injury (14, 15). These data, as well as the data presented here, allowed us to hypothesize that the Mincle may be an important sensor of nonphysiological cell death or cell stress that recognizes endogenous molecules released from damaged tissues and subsequently induces inflammation. The ribosomal protein SAP 130 has been reported to be an endogenous ligand for Mincle and to induce inflammation in vivo (7). We hypothesized that unknown endogenous lipid ligands for the Mincle receptor exist because the molecules of bacterial and fungal origin recognized by Mincle are amphiphilic lipids.

Isolation and Identification of Endogenous Mincle Ligands.

To evaluate whether Mincle recognizes endogenous lipids, we isolated the total lipid extract from mouse skin and tested its ability to induce the Mincle-dependent transcriptional activation of NF-κB in an NF-κB–driven luciferase reporter assay using a previously developed HEK293–mMincle–NF-κB–Luc cell line. As shown in Fig. 2A, the total lipid extract exhibited dose-dependent Mincle-stimulating activity. Next we performed a preliminary separation of the total lipid extract using a solid-phase extraction procedure. Using a reversed-phase cartridge, we found that the active compounds could be eluted from the cartridge almost without loss in their activating effect on Mincle (Fig. 2A). The chromatographic retention behavior suggested that the active components of the total lipid extract were amphiphilic molecules with both hydrophilic and hydrophobic fragments. Using electrospray ionization mass spectrometry, we found, perhaps expectedly, amphiphilic lipids in the eluate, particularly, cholesterol sulfate and phosphoinositols (Fig. S2). The extract obtained from the reversed-phase cartridge was further separated via HPLC in gradient separation mode using a reversed-phase monolithic column. Fractions were collected, and each fraction was evaluated using the reporter HEK293–mMincle–NF-κB–Luc cell line. The highest activity was detected in fraction 7 (Fig. 2B). One of the most intense signals in the mass spectrum (negative ion detection mode) of fraction 7 corresponded to a peak with an m/z of 465.31 (Fig. 2C). To elucidate the structure of this compound, this ion was isolated and subjected to fragmentation. Along with a peak corresponding to the parent ion in the tandem mass spectrum, only two product ion peaks were detected (m/z = 79.97 and m/z = 96.98, Fig. 2C). This finding suggests the presence of a sulfo-group in this compound and that the peak with an m/z of 465.31 corresponds to a deprotonated molecular ion [M-H]−. According to the LIPID MAPS database and MS-LAMP software, the compound was identified as a cholesterol sulfate. To confirm the identification, synthetic cholesterol sulfate was used as a standard. Its fragmentation spectrum completely matched the fragmentation spectrum of the ion from fraction 7 that had a peak with an m/z of 465.31 (Fig. 2E). Taken together, these data show that the main active component of fraction 7 was cholesterol sulfate.

Mincle Recognizes Cholesterol Sulfate Through Direct Binding and Induces a Mincle-Dependent Proinflammatory Response Both in Vitro and in Vivo.

First we performed surface plasmon resonance experiments to study the ability of cholesterol sulfate to bind with Mincle using highly purified synthetic cholesterol sulfate. An anti-histidine antibody was first covalently immobilized onto the sensor chip surface, and the histidine-tagged recombinant Mincle protein was then injected and captured by the immobilized anti-histidine antibody, after which the analytes were injected. To evaluate the specificity of the interaction, we used a negative control, synthetic 18:0–20:4 phosphoinositol. The lipid was coeluted from a reversed-phase cartridge together with cholesterol sulfate during a solid-phase extraction of skin lipids (Fig. S2), and it shared some physicochemical properties with cholesterol sulfate. We observed the direct binding of Mincle to cholesterol sulfate but not to phosphoinositol. The obtained sensorgrams are presented in Fig. 3A.

Next, we evaluated the ability of cholesterol sulfate to induce Mincle-dependent signaling using the reporter HEK293–mMincle–NF-κB–Luc cell line. The reporter cells showed a similar response to both cholesterol sulfate and trehalose dibehenate, an archetypical ligand of Mincle (Fig. 3B). Thus, this finding indicates that cholesterol sulfate binds to Mincle and induces Mincle-mediated signaling.

To determine the specificity of cholesterol sulfate signaling through Mincle under more physiologically realistic conditions and to study the relevance of the cholesterol sulfate/Mincle interaction in vivo, we used Mincle-deficient mice. To compare the ability of cholesterol sulfate and trehalose dibehenate to induce cytokine and chemokine secretion, bone marrow-derived dendritic cells (BMDCs) were isolated from Mincle-KO mice and wild-type mice and stimulated by these lipids. The levels of cytokines and chemokines in culture supernatants were measured using a multiplex immunoassay. Both cholesterol sulfate and trehalose dibehenate induced the secretion of proinflammatory mediators, such as IL-1a and IL-1b, KC, and MIP-1a and MIP-1b (Fig. 3C). The secretion of these proinflammatory mediators in response to both cholesterol sulfate and trehalose dibehenate was significantly reduced in BMDCs from Mincle-deficient mice, whereas the response to bacterial lipopolysaccharide (an agonist of TLR4) remained unchanged (Fig. 3C).

To evaluate the tissue response to cholesterol sulfate and the contribution of Mincle to this process, we subcutaneously injected a sterile aqueous solution containing micelles of cholesterol sulfate into either Mincle-deficient or wild-type mice. Sterols tend to form crystals in solution, and these crystals are potent activators of inflammatory responses (16, 17). To avoid the injection of a crystal-containing suspension, we previously developed a method using an ultrasonic treatment to produce a stable micelle solution with an average micelle diameter from 70 to 90 nm (Fig. S3). After 24 h, skin samples were excised and the local tissue response was evaluated based on a histological analysis of hematoxylin and eosin (H&E)-stained sections. The histological examination revealed a marked inflammatory infiltrate in the skin of wild-type animals injected with cholesterol sulfate, whereas, in Mincle-deficient animals, the accumulation of infiltrating cells was significantly reduced (Fig. 3D). The infiltrating cells were primarily neutrophils, monocytes, and eosinophils. Taken together, our findings suggest that a Mincle deficiency results in a reduced inflammatory response to cholesterol sulfate both in vitro and in vivo.

Mincle Promotes Cutaneous Allergic Inflammation.

Taking into account (i) the significant inducibility of Mincle expression in skin to the topical skin irritation caused by DNFB, a well-known inductor of allergic skin reactions, and (ii) the significantly impaired secretion of proinflammatory mediators in response to a single DNFB application in Mincle-deficient animals, we examined the role of Mincle in the pathogenesis of cutaneous allergic inflammation. To address the role of Mincle in the development of cutaneous allergic inflammation, we used one of the most commonly used models of contact hypersensitivity (CHS) that adequately reflect allergic contact dermatitis (18). Mincle-KO and wild-type mice were sensitized to DNFB on their abdomens, followed by challenges with the same sensitizer on the dorsum of both ears. The schedule for the mouse sensitization and challenge procedure is depicted in Fig. 4A. Endpoint parameters were measured 24 h after the last DNFB exposure. Under the induction protocol, Mincle-deficient mice exhibited significantly suppressed clinical symptoms of allergic contact dermatitis, such as redness and ear swelling, compared with the wild-type animals (Fig. 4 B and C). A careful review of H&E-stained stained sections of contact-allergic ear tissue at the experiment endpoint revealed a strongly reduced inflammatory response in Mincle-deficient mice in terms of ear thickness, epidermal acanthosis, and inflammatory cell infiltrate (Fig. 4D). A quantitative analysis using flow cytometry confirmed that there were significantly fewer infiltrating Gr1+/CD11b+ myeloid cells in the ear tissue of Mincle-deficient mice compared with wild-type control mice (Fig. 4E). To evaluate the differences in allergic skin inflammation at the molecular level, we measured the production of cytokines and chemokines in inflamed ear tissues. We found that a Mincle deficiency led to the suppressed production of a spectrum of mediators of inflammation (Fig. 4F), including proinflammatory cytokines IL-1b, TNFa, and IL12p40, as well as chemokines KC, MIP-1a, and MIP-1b. Interestingly, the production of IL-17A was also significantly suppressed in the inflamed tissues of Mincle-deficient mice compared with IL-17A’s production in wild-type animals (Fig. 4F). Taken together, these observations demonstrate that Mincle strongly enhances the magnitude of the skin inflammatory CHS response and plays a role as a driving component in the pathogenesis of allergic skin inflammation.

Mice.

Male BALB/c mice (Stolbovaya Nursery of the Russian Academy of Medical Science) and C57BL/6 and C57BL/6-Mincle-KO mice of both sexes were used in this study. All mice were between 9 wk and 10 wk of age. The Mincle-KO knockout mouse line was obtained from the National Institutes of Health-sponsored MMRRC National System and was back-crossed with C57BL/6 mice background for 10 generations. The mice were fed a completely pelleted laboratory chow and had access to food and water ad libitum.

Wounding Procedure.

Male BALB/c mice (for microarray analysis) or male C57BL/6 and C57BL/6-Mincle-KO mice were used. The mice were anesthetized intraperitoneally with xylazine (7.5 mg/kg) and Zoletil (45 mg/kg, Virbac). The backs of the mice were shaved with an animal clipper and disinfected with 70% (vol/vol) ethanol. Two full-thickness dermal wounds were made on opposite sides of the midline of each mouse using a 4-mm punch biopsy instrument (Dermo-Punch). Wounds were made through the epidermis, dermis, and s.c. tissue layer, leaving the fascia intact. The animals were individually caged after wounding. At a specified time point after wounding, mice were killed via carbon dioxide inhalation. Then the wounds and surrounding tissues or intact skin samples were removed with an 8-mm biopsy punch.

In Vivo Wound-Healing Assay.

Digital photographs of the wounds were taken on days 1, 3, 7, and 12 after wounding. The area of the wound was then measured on the photographs in pixels using Adobe Photoshop CS6 (Adobe).

Microarray Analysis.

Tissue samples were crushed in liquid nitrogen. QIAzol lysis reagent (Qiagen) was added, and the homogenate was centrifuged through a QIAshredder column (Qiagen). Total RNA was extracted from the eluate using the miRNeasy mini kit (Qiagen) according to the manufacturer’s instructions. Genomic DNA contamination was removed by performing a DNaseI digestion on an RNA-binding column for 15 min. Total RNA was eluted in 50 μL of RNase-free water and stored at −80 °C. RNA yield and purity were assessed spectrophotometrically by measuring OD₂₆₀ and the OD_260/280 ratio, respectively, in RNase-free water using a NanoDrop ND-1000 (NanoDrop Technologies). For all samples, the OD_260/280 ratio was between 2.0 and 2.2.

RNA integrity was determined using an Agilent Bioanalyzer 2100 system (Agilent Technologies). RNA quality indicator (RIN, RNA integrity number) values generated by the Bioanalyzer system’s software ranged from 7.7 to 9.1.

RNA samples were prepared according to the manufacturer’s instructions (Affymetrix Manual P/N 701880 Rev. 4) using 500 ng of total RNA as the starting material as described previously (43, 44). The samples were hybridized on GeneChip Mouse Gene 1.0 ST Arrays (Affymetrix) for 16 h at 45 °C. Arrays were washed to remove nonspecifically bound nucleic acids and stained on a Fluidics Station 450 (Affymetrix) using the FS450_0007 protocol followed by scanning on a GeneChip Scanner 3000 7G system (Affymetrix). The microarray CEL files have been deposited in the GEO database (accession no. GSE76144).

Arrays were preprocessed in GeneChip Command Console Software (AGCC, Affymetrix, version 1.4.1.46) and the Affymetrix Transcriptome Analysis Console (version 3.0.0.466) using the RMA-sketch option. Differentially expressed genes were determined based on a moderated t test using the limma package from the Bioconductor project (www.bioconductor.org). Genes with absolute fold changes above 3.0 and FDR-adjusted P values below 0.05 (Benjamini–Hochberg procedure) were selected.

Western Blot Analysis.

Skin samples containing wounds or sites of DNFB or vehicle application as well as intact skin were removed using an 8-mm biopsy punch and then immediately placed into ice-cold T-PER extraction buffer (T-PER, Pierce) containing a complete protease inhibitor (Roche Diagnostics). Approximately 0.75 mL of the buffer was used per 35 mg of skin sample. The homogenized samples were prepared using a FastPrep 24 device and tubes containing lysing matrix A (MP Biomedicals). The homogenates were centrifuged at 12,000 × g for 12 min at 4 °C. The total protein in the supernatants was measured using the Bradford method, and protein levels in the supernatants were then normalized. Next the protein extracts were separated via SDS/PAGE, transferred onto nitrocellulose membranes, and probed with antibodies against Clec4e (1:1,000, D292-3, MBL International) or β-actin (1:3,000, ab8227, Abcam). Proteins of interest were detected with an HRP-conjugated goat anti-rat IgG antibody (1:10,000, NA935, GE Healthcare) or goat anti-rabbit IgG antibody (1:2,000, ab6721, Abcam) and visualized using the Optiblot ECL detection kit (ab133406, Abcam) according to the provided protocol.

Immunohistochemistry.

Skin samples containing wounds or sites of DNFB or vehicle application as well as intact skin were removed using an 8-mm biopsy punch and frozen in liquid nitrogen. Next we cut 4-μm sections of the skin samples and fixed them using 4% (vol/vol) paraformaldehyde. The samples were blocked with BSA (Sigma-Aldrich) for 30 min at 37 °C. They were then incubated with primary antibodies for 1 h at 37 °C. The following primary antibodies were used: anti-Clec4e (1:200, orb1341, Biorbyt), anti-CD45R/B220 (1:200, 103201, BioLegend), anti-BST2 (120G8.04) (1:10, DDX0390P, Novus Bio), anti-CD11c (117302, BioLegend), and anti–Langerin-PE (1:200, 144204, BioLegend). After washing, we incubated the samples with secondary antibodies for 30 min at 37 °C. Next, the cell nuclei were costained with DAPI (Sigma-Aldrich). The following secondary antibodies conjugated with fluorochromes were used: anti-rabbit antibodies (1:100, 711–585-152, Jackson ImmunoResearch), anti-rat antibodies (1:500, 10123952, Invitrogen), and anti-Armenian hamster antibody (1:50, 405512, BioLegend). The samples were photographed using a Keyence microscope (Model BZ-9000) and a confocal microscope (TCS SP5 STED; Leica Microsystems).

Extraction of Total Lipids from Skin.

Intact murine skin samples were collected, weighed, cut with scissors, and then immediately placed into ice-cold RIPA buffer. Approximately 7.5 mL of the buffer was used per 1 g of skin. The homogenized samples were prepared using a FastPrep 24 device and tubes containing lysing matrix A (MP Biomedicals). Total lipids were isolated from the obtained homogenate using sequential extractions with different solvent mixtures according to a procedure described previously (45). The obtained total lipid extract was dried under a vacuum.

Solid-Phase Extraction.

Approximately 25 mg of the obtained total lipid extract was emulsified in 10 mL of a chloroform–methanol–water (3/48/47) mixture containing 0.1 M NaCl, and the emulsion was then sonicated using an ultrasound bath. The lipids were fractionated according to a previously reported procedure (46). Briefly, the lipid extract was applied to a pre-equilibrated Sep-Pak C18 cartridge (WAT051910, Waters) connected to an AKTA Explorer 10 chromatography system (GE Healthcare), and the cartridge was then sequentially washed at a flow rate of 3 mL/min with a chloroform–methanol–water (3/48/47) mixture containing 0.1 M NaCl, followed by water and, finally, methanol. The methanol-eluted fraction was then collected and dried using a vacuum rotary evaporator.

Separation via HPLC.

HPLC separations were achieved as described by Ikeda et al. (47) with several important changes. Briefly, HPLC was performed using a monolithic reversed-phase Chromolith Performance RP-18e column (4.6 mm × 100 mm, Merck). Separations were performed using a Waters BioAlliance 2796 HPLC module (Waters). The column temperature was held at 40 °C. The mobile phases were prepared from mixtures of methanol (A), isopropanol (B), and water (containing 200 mM ammonium formate) (C). The gradient consisted of holding solvent D (A/B/C, 55/25/20 by volume) for 5 min, linearly converting solvent D to solvent E (A/B, 60/40 by volume) over 10 min, and then holding solvent E for 5 min. After a run was completed, the gradient was returned to solvent D for over 5 min, and before the next run, an equilibration of the column was performed for 5 min. The mobile phase was pumped at a flow rate of 1.5 mL/min for gradient elution. Approximately 1 mg of obtained, dried eluate from the solid-phase extraction was dissolved in solvent D. After sonication in an ultrasound bath, 100 μL of the sample solution was applied to the column, and 1-mL fractions were collected using a fraction collector. Fractions of the same number from three independent runs were combined and then dried using a vacuum rotary evaporator.

Lipid Identification via Electrospray Ionization Mass Spectrometry Analysis.

The electrospray ionization mass spectrometry (ESI-MS) analyses were performed using a QSTAR Elite system (AB Sciex) equipped with a TurboIonSpray ion source via direct infusion using a built-in Harvard Syringe Pump at a flow rate of 5 µL/min. Full-scan profiling data were acquired in the TOF mass analyzer in positive and negative modes, and the source and ion transfer parameters applied were as follows: ion source nebulizer gas (GS1), 3 L/min; ion source heater gas (GS2), 3 L/min; gas temperature (TEM), 300 °C; and curtain gas flow rate (CUR), 20 L/min. The ionspray voltage (IS) was −4,500 V in negative mode and 5,500 V in positive mode. The declustering potential (DP) was −70 V, and the focusing potential (FP) was −40 V. Tandem mass-spectrometry experiments (MS/MS) were performed using the same parameters for the ion source and optics as described above. Parent ions were isolated in quadrupole Q1 in the “Unit Resolution” mode. The collision energy for ion fragmentation was set to 40 eV.

An identification search was performed in the LIPID MAPS database (www.lipidmaps.org) with the help of MS-LAMP software (ms-lamp.igib.res.in) using the following queried parameters: m/z, 465.3113 (monoisotopic); ion type, [M-H]−; window range, ±0.01; class, sulfur-containing lipids. As a result of this query, only one possible molecule, cholesterol sulfate, was proposed; thus, the precision of the mass measurement was 0.0074 Da or 16 ppm.

Synthetic cholesterol 3-sulfate (700016P, Avanti Polar Lipids) was used as a standard.

Lipid Preparation.

A micelle-containing solution of cholesterol 3-sulfate (700016P, Avanti Polar Lipids) was prepared at a concentration of 1 mg/mL for in vivo treatments using a Branson S-450D instrument (Branson). Endotoxin-free water was obtained from a Milli-Q Advantage A10 system (Millipore). The ultrasonic treatment produced a solution of micelles that were stable for at least 1 h (according to the size and size distribution measurements, Fig. S3), with the average micelle diameter ranging from 70 nm to 90 nm. Before injection, the solution was sterilized via filtration through a 22-µm filter. Measurements of cholesterol sulfate micelle size and size distribution were performed in triplicate using a Zetasizer Nano ZS (ZEN3600, Malvern Instruments Ltd.) at 25 °C. All calculations based on light-scattering intensity were performed using Zetasizer Nano ZS software.

To perform the Surface Plasmon Resonance (SPR) experiments and the reporter cell-line assays, synthetic lipids (trehalose dibehenate, 890808; cholesterol sulfate, 700016P; 18:0–20:4 phosphoinositol, 850144P; all from Avanti Polar Lipids) as well as dried total lipid extract aliquots and separated fractions were solubilized using the ultrasound treatment described above. Endotoxin-free water was used as a solvent in the reporter cell-line assays, and binding buffer was used in the SPR experiments. Then serial dilutions were performed, and each dilution was sonicated.

For experiments with BMDCs, 1 mg of cholesterol sulfate (700016P, Avanti Polar Lipids) or trehalose dibehenate (vac-tdb, InvivoGen) was dissolved in 100 µL of DMSO (DMSO Hybri-Max, Sigma-Aldrich) and heated at 60 °C for 30 s. Then 900 µL of water was added, followed by heating for 15 min at 60 °C, after which serial dilutions were performed. In in vitro experiments, bacterial lipopolysaccharide from Escherichia coli O111:B4 (tlrl-pelps, InvivoGen) and recombinant human TNFa (rhtnf-a, InvivoGen) were used as controls.

SPR-Binding Experiment.

The SPR experiments were performed using a BIACORE 3000 (GE Healthcare) equipped with a research-grade CM5 sensor chip (BR100012, GE Healthcare) at a temperature of 25 °C. For binding analyses, the anti-histidine antibody from the His Capture Kit (28-9950-56, GE Healthcare) was first covalently immobilized onto the sensor chip surface at a level of ∼15,000 response units (RU) using the Amine Coupling Kit (BR-1000-50). Histidine-tagged recombinant human Clec4e, derived from human cells (C588, Novoprotein) in running buffer (25 mM Tris⋅HCl, 150 mM NaCl, 2 mM CaCl₂, pH 7.2), was then injected over flow cell 2 (Fc2) at a concentration of 10 μg/mL at a flow rate of 15 µL/min and captured by the immobilized anti-histidine antibody. Flow cell 1 (Fc1), with immobilized anti-histidine antibodies, was intact and used as a blank surface. The analytes cholesterol 3-sulfate, sodium salt (700016P, Avanti Polar Lipids), and 18:0–20:4 phosphoinositol (1-stearoyl-2-arachidonoyl-sn-glycero-3-phosphoinositol, ammonium salt, 850144P, Avanti Polar Lipids) were prepared in running buffer using the ultrasound treatment described above. The analytes were injected over the two flow cells (Fc1 and Fc2) at a flow rate of 20 μL/min, and the complex was then allowed to associate and dissociate. Then the surfaces were regenerated with a 30-s injection of the regeneration solution (10 mM Tris⋅glycine, pH 1.5). During regeneration, the captured histidine-tagged ligand and the associated analyte were removed from the sensor chip surfaces, making it ready for the next cycle. Injections of each analyte concentration and a buffer blank were flowed over the two flow cells. Data were collected at a rate of 1 Hz. The data were double-referenced (blank surface and blank buffer referencing) using BIAevaluation software (GE Healthcare).

Reporter Cell-Line Preparation.

Mouse Clec4e (mMincle, gene ID 56619) and mouse Clec4d (mMCL, gene ID 17474) cDNA sequences were synthesized by Evrogen and inserted into pAI-TA2 vectors (Evrogen). For lentiviral stock preparation, Clec4e and Clec4d gene nucleotide sequences were cloned in the lentiviral plasmids pLV-Neo and pLV-Bleo, respectively. For the preparation of lentiviral particles, 293T cells (ATCC, CRL-3216) were transfected with vectors coding for Clec4e (pLV-Clec4e-Neo) or Clec4d (pLV-Clec4d-Bleo), together with the NF-κB–Luc construct (pLA-NF-κB–Luc), using Lipofectamine reagent (Invitrogen) according to the manufacturer’s instructions. Two helper plasmids, pCMVΔR8.2 and pVSV-G, were cotransfected along with the experimental vector. Supernatants containing infectious viral particles were harvested 24, 36, and 48 h post transfection, pooled, and filtered through a 0.22-µm filter.

HEK293 cells (ATCC, CRL-1573) were transduced with the lentiviral vector containing the reporter NF-κB–Luc construct pLA-NF-κB–Luc and were then cultured in the presence of 1 µg/mL puromycin (Invitrogen) for 2 d to produce a control reporter cell line (HEK293–NF-κB–Luc). Next, the control cells were transduced with either Mincle- or MCL-coding lentiviral vectors and cultured in 500 µg/mL G418 sulfate and 10 µg/mL bleomycin (both from Invitrogen) for 1 wk to produce stable cell lines (HEK–mMincle–NF-κB–Luc and HEK–mMCL–NF-κB–Luc). The expression of Mincle and MCL in the reporter cells was validated via Western blotting using anti-Mincle antibodies (1:1,000, D292-3, MBL International) or anti-MCL antibodies (1:200, sc-134743, Santa Cruz Biotechnology). Functional testing was performed by assessing the nuclear translocation of NF-κB after stimulating the cells with trehalose dibehenate. Western blotting of nuclear fractions was then performed using anti-p65 antibodies (1:2,000, ab16502, Abcam) as described elsewhere.

The reporter cells were maintained in DMEM (PAA Laboratories) supplemented with 10% (vol/vol) FCS (Thermo Scientific), 50 U/mL penicillin, 50 μg/mL streptomycin, 2 mM glutamine, and 0.1 M NaHCO₃ (all from PanEco) at 37 °C with 5% (vol/vol) CO₂.

Reporter Cell-Line Assay.

HEK–mMincle–NF-κB–Luc or HEK–NF-κB–Luc cells were seeded at 2 × 10⁴ cells/well in 96-well plates (100 µL/well). Cells were treated with the substances of interest for 8 h. After incubation, the cells were lysed by adding 100 µL per well of GloLysis buffer (Promega) according to the manufacturer’s instructions. Luciferase activity was measured in the cell lysates using the Bright-Glo luciferase assay system (Promega). Luminescence was measured in relative units using a Synergy H4 hybrid reader (Bio-Tek Instruments).

BMDC Preparation.

BMDCs from C57BL/6 and C57BL/6-Mincle-KO knockout mice were differentiated from proliferating mouse bone marrow progenitor cells via induction with 20 ng/mL of granulocyte macrophage colony-stimulating factor (GM-CSF) (R&D Systems) over 6–9 d as described previously (48). Briefly, mice were euthanized with carbon dioxide. Their femurs and tibias were collected in ice-cold HBSS (HBSS, Sigma-Aldrich). The muscles were removed with a scalpel and by rubbing the bones with a paper tissue. The ends of the bones were then cut off with scissors and crushed. The bone marrow was flushed out with 2–3 mL of RPMI complete medium in a syringe with a 25-gauge needle. All bone marrow cells were collected and washed twice with HBSS. The bone marrow cells were cultured in 24-well plates containing ∼5 × 10⁵ cells/mL in 1 mL total volume. The cells were maintained at 37 °C with 5% (vol/vol) CO₂ in complete RPMI medium with 10% (vol/vol) heat-inactivated FCS (PAA Laboratories), 0.05 M mercaptoethanol (Thermo Fisher Scientific), nonessential amino acids (PanEco), 20 ng/mL GM-CSF, 2 mM glutamine, 100 U/mL penicillin, and 100 g/mL streptomycin (all PanEco). After 3 d, the nonadherent cells were collected and discarded and 1 mL of fresh medium was added to each well. On day 5, half of the medium was replaced with fresh medium. On day 7, the cells were collected, washed, and resuspended in medium without GM-CSF. After 7–9 d of cultivation, the percentage of CD11c+ cells was ∼60–70%. BMDCs obtained from C57BL/6 and C57BL/6-Mincle-KO knockout mice were seeded in 96-well plates at 1 × 10⁵ cells per well (200 µL/well) in complete RPMI medium. Cells were treated with the substances of interest for 48 h. Plates were then centrifuged at 1,000 × g for 10 min, and the culture supernatants were collected.

In Vivo Cholesterol Sulfate Treatments.

C57BL/6 and C57BL/6-Mincle-KO knockout mice were anesthetized intraperitoneally with xylazine (7.5 mg/kg) and Zoletil (45 mg/kg, Virbac), and the backs of the mice were shaved with an animal clipper. After disinfection with 70% (vol/vol) ethanol, mice were injected s.c. with 20 µL of cholesterol sulfate solution or just the solvent (sterile water) as a negative control. Twenty-four hours post injection, the mice were anesthetized, and skin samples from the injection site were removed with an 8-mm biopsy punch.

Induction of Allergic Contact Dermatitis.

Allergic contact dermatitis was induced using a method described previously (18) with minor modifications. Briefly, Mincle-KO and wild-type mice were sensitized with 25 μL of 0.5% DNFB (Sigma-Aldrich) on the shaved abdomen for 2 consecutive days (days 0 and 1). Immediately before application DNFB was dissolved in vehicle (acetone (AppliChem) and olive oil (Sigma-Aldrich) in a proportion of 4:1 by volume). Challenges with 10 μL of 0.3% DNFB were performed on the dorsum of both ears on days 5, 6, 7, and 8, and the endpoint parameters were measured 24 h after the last DNFB exposure. As a negative control, two other groups of Mincle-KO and wild-type mice were exposed to the vehicle without DNFB for the duration of the experiment. Ear thickness was measured with a digital micrometer (Mitutoyo).

Cytokine and Chemokine Assays.

Skin samples containing wounds or sites of DNFB or vehicle application as well as intact skin were removed using an 8-mm biopsy punch. The skin samples or ears were immediately placed into ice-cold T-PER extraction buffer (T-PER, Pierce) containing a complete protease inhibitor (Roche Diagnostics). Approximately 0.75 mL of the buffer was used for each skin sample or ear. The homogenized samples were prepared using a FastPrep 24 device and tubes containing lysing matrix A (MP Biomedicals). The homogenates were centrifuged at 12,000 × g for 12 min at 4 °C. The total protein in the supernatant was measured using the Bradford method.

Levels of 23 cytokines and chemokines [IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12 (p40), IL-12 (p70), IL-13, IL-17A, eotaxin (CCL11), G-CSF, GM-CSF, IFN-γ, KC (CXCL1), MCP-1 (CCL2), MIP-1α (CCL3), MIP-1β (CCL4), RANTES (CCL5), TNF-α] were measured in triplicate in the prepared cell culture supernatants and in the tissue homogenates using a 23-plex ELISA Bio-Plex Pro kit (Bio-Rad Laboratories) according to the manufacturer’s instructions. The concentrations of some selected cytokines and chemokines in cell culture supernatants were validated using commercial ELISA kits according to the manufacturers’ instructions. ELISA kits for KC, IL-1β, and MIP-1β were purchased from R&D Systems; ELISA kits for TNF-α and IL-1α were purchased from eBioscience; and ELISA kits for MIP-1α were purchased from Abcam.

Immunophenotyping of Ear Cells.

Ears from each mouse were dissected, disrupted mechanically using scissors, and placed in 1 mL of PBS solution containing 5% (vol/vol) FBS (Thermo Fisher Scientific) and 1 mg/mL collagenase type IV (Sigma-Aldrich). After 30 min of incubation at 37 °C, cells were centrifuged at 400 × g for 10 min, collected, and washed twice with PBS before flow cytometric analysis. Staining was performed with the fluorochrome-conjugated anti-CD11b Alexa Fluor 700 (clone M1/70), anti–Ly-6G and anti–Ly-6C (Gr1) FITC (clone RB6-8C5) monoclonal antibodies, or the corresponding isotype controls for 20 min at 4 °C in Staining Buffer (all BD Biosciences). Fluorescence was measured using a FACSAria III flow cytometer system, and the data were analyzed using FACSDiva software (all BD Biosciences).

Histological Evaluations.

Skin samples and ears were fixed in modified Davidson’s fluid and processed for histological analysis. Tissue blocks were embedded in paraffin. Series of 4-µm sections were cut through each 100 µm of tissue, and the sections were then stained with Caracci’s hematoxylin and eosin. All sections were evaluated by two pathologists blinded to the experiment’s design. For each animal, slides with the maximum inflammatory response score were selected. Microscopic images were obtained using a Keyence microscope (Model BZ-9000).

Cholesterol Sulfate Quantification in Skin.

Skin samples containing sites of DNFB or vehicle application as well as intact skin were removed using an 8-mm biopsy punch and weighed. Next the skin samples were immediately placed into tubes containing lysing matrix A (MP Biomedicals) and 0.75 mL of extraction mixture (chloroform/methanol/water, 1/2/0.8, by volume). The homogenized samples were prepared using a FastPrep 24 device and tubes containing lysing matrix A (MP Biomedicals). The homogenates were centrifuged twice at 12,000 × g for 10 min, and the supernatants were used for cholesterol sulfate quantification.

HPLC separations and cholesterol sulfate detection were conducted as described previously in the literature (49) with several important modifications. Briefly, HPLC was performed using a reversed-phase Zorbax Eclipse XBD C18, 2.1 × 150 mm, 5 µm (Agilent). The separations were performed using a Waters BioAlliance 2796 HPLC module (Waters) at room temperature with a gradient of mobile phase A [100 µM ammonium formate in 50% (vol/vol) methanol with 0.1% formic acid] and mobile phase B [100% (vol/vol) methanol]. The flow rate was 200 μL/min using gradient elution: 0–2 min, 60% B; 2–4 min, 60–90% B; 4–20 min, 90% B. The injection volume was 10 µL. The ESI-MS analyses were performed using a QSTAR Elite system (AB Sciex) equipped with a TurboIonSpray ion source. Full-scan profiling data were acquired in the TOF mass analyzer in negative ion mode, and we used the following source and ion transfer parameters: ion source nebulizer gas (GS1), 10 l/min; ion source heater gas (GS2), 5 L/min; TEM, 425 °C; and CUR, 20 L/min. The IS was −4,500 V. The DP was −70 V, and the FP was −260 V. The mass spectrometer was operated using Analyst software version 2.0 (AB Sciex). The quantitation analysis was performed in “precusor ion scan mode” in negative ion mode. In this scan type, quadrupole Q1 was scanned in the mass range of 460–470 Thomson (Th), and the ions were fragmented in the Q2 collusion cell. The ions produced were pulsed into the TOF and measured at the detector over a mass range of 50–600 Th. We used diagnostic fragment ions of 96.9700 Da and 79.9700 Da (center of precursor masses) and a mass window of 2 Da.

As a standard for quantification, we used cholesterol 3-sulfate (700016P, Avanti Polar Lipids).

Statistical Analysis.

All experiments were repeated at least twice (except for the microarray analysis). Unless stated otherwise, the statistical significance of the differences among the means was determined using a one-way ANOVA with the Statistica 7 statistical software package (StatSoft, Inc.). Differences were considered significant if P was less than 0.05.