Role of dendritic cell maturation factors produced by human invariant NKT cells in immune tolerance

Short abstract

S100A8 promotes T_reg generation by activity on modulating immature DCs in the presence of GM‐CSF and IL‐4.

Abstract

In this study, we used the culture supernatant of iNKT cells to identify human myeloid DC maturation factors produced by human CD4⁺ iNKT cells. S100A8 had a strong maturation effect. Notably, the recombinant S100A8 protein displayed properties of DC maturation functioning, and the induction of DC differentiation by both the purified and the recombinant protein were blocked by anti‐S100A8 and anti‐TLR‐4 mAbs. DC differentiation induced by anti‐major histocompatibility complex class II/CD1d Ab, S100A8, or both was qualitatively indistinguishable from that induced by the coculture of DCs and iNKT cells or via culture supplementation with supernatants from activated CD4⁺ iNKT cells. S100A8 also induced CD4⁺/CD25⁺/Foxp3⁺ T_reg cells from naïve T cells. S100A8 may contribute to DC differentiation by elevating transcription factors or activating transcription factor‐2, heat shock factor‐1, or both, in mature DCs. S100A8 is a novel candidate iNKT cell‐dependent DC maturation factor.

Abbreviations

α‐GalCer

α‐galactosylceramide

ATF

activating transcription factor

BMDC

bone marrow‐derived dendritic cell

CDR

complementarity determining region

CFSE

carboxyfluorescein diacetate succinimidyl ester

DC

dendritic cell

DN

double negative

FPLC

fast protein liquid chromatography

HIF

hypoxia inducible factor

Hsp70

heat shock protein 70

iDC

immature dendritic cell

iNKT cell

invariant natural killer T cell

PBMC

peripheral blood mononuclear cell

PDL

programmed death ligand

PIP

prolactin‐induced protein

RAGE

receptor for advanced glycation end products

S100A8

S100 calcium‐binding protein A8

S100A9

S100 calcium‐binding protein A9

siRNA

small interfering ribonucleic acid

T_reg

regulatory T cell

Introduction

Human iNKT cells show restricted use of the invariant Vα24‐Jα15 TCR and express several cell surface proteins found on NK cells [1]. Therefore, iNKT cells share functional and phenotypic homology with both NK and T cells and are distinguished by the nonpolymorphic class Ib molecule CD1d, which is involved in the presentation of glycolipid antigen [2]. This presentation results in a burst of secretion of several cytokines—particularly, IL‐4 and IFN‐γ [3]. The rapid secretion of cytokines implies that iNKT cells play a key role in the modulation of immune response and immunopathology. iNKT cells are also involved in tumor immunity [4], diabetes regulation [5, 6–7], and protection against bacterial [8], viral [9], and parasitic infections [10].

iNKT cells have also been shown to promote peripheral tolerance in several model systems, but their regulatory effects remain poorly understood. The soluble factors secreted by human iNKT cells instruct human PBMCs to differentiate into myeloid APCs with suppressive properties. In addition, human iNKT cells direct primary human PBMCs to differentiate into cells resembling immature myeloid DCs [11]. iNKT cell activation via the recognition of CD1d expressed on monocytes results in the secretion of GM‐CSF and IL‐13, which promotes monocyte differentiation. The resulting myeloid cells show upregulation of DC‐specific intercellular adhesion molecule‐3‐grabbing nonintegrin, little or no expression of CD14, and moderate expression levels of the costimulatory markers CD40 and CD86. Thus, the cell surface phenotype is consistent with that of immature myeloid DCs. The localization of MHC class II molecules to lysosomal‐associated membrane protein 1‐positive intracellular vesicles, which is characteristic of immature myeloid DCs, was also observed. After exposure to LPS, immature myeloid DCs undergo changes associated with maturation, including the upregulation of CD83 and chemokine receptor 7 and the relocation of MHC class II molecules to the cell surface [11]. DCs fail to mature normally in both human and rodent autoimmune diabetes.

The transfer of highly mature DCs protects nonobese diabetic mice from disease. Therefore, defects in APCs and CD1d‐restricted T cells may underlie the development of pathogenic autoimmune T cells and type 1 diabetes. Based on our observations, the natural history of type 1 diabetes progression is influenced by a fine balance in which a T_h 1‐like response is critical for β‐cell destruction and progression toward overt disease. If the autoimmune T‐cell pool becomes biased toward T_h2‐like cells (e.g., those that can inhibit T_h1 cells via IL‐4 or ‐10), either naturally or via the activation or transfer of the appropriate CD1d‐restricted T‐cell subset, the disease process can be halted.

S100A8 (encoded by the Mrp8 gene) and S100A9 (encoded by the Mrp14 gene) are the most abundant cytoplasmic proteins in neutrophils and monocytes [12]. S100A8 and S100A9 belong to the calcium‐binding S100 protein family, and S100A8 forms Ca²⁺‐dependent heterodimer/heterotetramer complexes with S100A9. Human S100A8 is 93 aa (10.8 kDa) and contains 2 EF‐hand motifs (aa 12–47 and aa 46–81) and 1 high‐affinity Ca²⁺ binding site (aa 59–70). Human S100A9 is also an EF‐hand calcium‐binding protein. It consists of 114 aa (13.2 kDa) and is expressed in granulocytes, monocytes, and macrophages during acute and chronic inflammation. S100A8 is detected in epithelial cells during dermatoses and has antimicrobial activity against bacteria and fungi. This molecule plays a role in the development of endotoxic shock in response to bacterial LPS. In addition, S100A8 functions as a proinflammatory mediator during acute and chronic inflammation and upregulates the release of IL‐8 and the cell surface expression of ICAM‐1 on the endothelium.

S100A8 shares a 57% amino acid identity with mouse S100A8 [12, 13]. A recent report showed that S100A8 is an endogenous activator of TLR‐4; thus, it elevates the expression of TNF‐α [14]. In the present study, mass spectrometric analysis of the supernatant from activated CD4⁺ iNKT cells demonstrated that S100A8 is induced during anti‐CD3 Ab or α‐GalCer activation. Furthermore, our results demonstrated that S100A8 induces the maturation of iDCs and generates T_reg cells.

MATERIALS AND METHODS

Cell lines

CD1d‐restricted T‐cell clones were generated via single‐cell sorting by using MoFlo (BD Biosciences, Mountain View, CA, USA). In brief, NKT cells were sorted by using 6B11‐fluorochrome conjugated Ab (an mAb specific for the invariant Vα24JαQ CDR 3 loop) [15], and single‐cell sorts were grown with a mixture of irradiated (5000 rad) allogeneic PBMCs at a density of 75,000 cells per well. The NKT cell clones were frozen in liquid nitrogen until further use. After thawing, the clones were expanded with 100 ng/ml α‐GalCer and γ‐irradiated PBMCs. These experiments were conducted with the informed consent of each participant and the approval of the Inha University ethics committee.

Culture of iNKT cell clones and transfection

iNKT cell clones were expanded via culture in RPMI 1640 (BioWhittaker, Walkersville, MD, USA) supplemented with 10% heat‐inactivated FBS (Atlanta Biologicals, Norcross, GA, USA), 25 × 10⁶ irradiated PBMCs, 100 ng/ml α‐GalCer, 2 mM l‐glutamine, 10 mM HEPES buffer, 100 U/ml penicillin (BioWhittaker), and 100 μg/ml streptomycin sulfate (BioWhittaker) [16]. Cells were incubated at 37°C in a humidified chamber with 5% CO₂. After 18–24 h, 50 U/ml human recombinant IL‐2 (Roche, Mannheim, Germany) and 10 U/ml human IL‐7 (Roche) were added to cocultured iNKT and feeder cells. On d 5, half of the medium was replaced with fresh medium supplemented with 50 U/ml IL‐2 and 10 U/ml IL‐7. During d 10–14, the iNKT cells were split for further expansion [17]. The purity of the expanded cells was checked with flow cytometry by using anti‐CD4, ‐CD8, and ‐6B11‐fluorochrome–conjugated Abs. Cell transfection with siRNA against S100A8 and PIP was performed with S100A8 and PIP Trilencer‐27 Human siRNA (OriGene, Rockville, MD, USA), according to the manufacturer's protocol. Cells were allowed to recover for 24 h before use.

Human cytokine Ab array

CD4⁺ and DN iNKT cells were stimulated with anti‐CD3 Ab, as described above. The supernatants were collected after 24 h and stored at −80°C. The conditioned medium was analyzed with a RayBio Human Cytokine Antibody Array C Series 1000 (RayBiotech, Norcross, GA, USA), according to the manufacturer's protocol. In brief, the membranes were incubated in blocking buffer for 30 min, followed by overnight incubation with conditioned medium at 4°C. The membranes were washed 5 times with washing buffer and incubated for 2 h with biotin‐conjugated Abs. The membranes were then washed 5 times with washing buffer and incubated for 2 h with HRP‐conjugated streptavidin. After the washing process, human cytokines were detected with enhanced chemiluminescence reagents.

RT‐PCR analysis

For RT‐PCR, mRNA was isolated with the RNeasy mini kit (Qiagen, Valencia, CA, USA), according to the manufacturer's instructions. Total RNA (1 μg), S100A8‐F (sense, ACC GAG CTG GAG AAA GCC TTG AAC TCT), and S100A8‐R (antisense, CTC TTT GTG GCT TTC ATG GCT TTT) primers and the RT Super Script II enzyme (Thermo Fisher Scientific, Carlsbad, CA, USA) were used for the RT‐PCR experiments. The first strand of complementary DNA was synthesized at 50°C for 30 min, and 34 cycles (94°C for 30 s, 55°C for 30 s, and 72°C for 60 s) were used to amplify the S100A8 gene, yielding a PCR product with an expected size of 273 bp.

Preparation of anti‐CD3‐activated NKT cell culture supernatant

For the preparation of supernatant from CD4⁺ iNKT cells activated with anti‐CD3 Ab, a plate was treated with 100 ng/ml of anti‐CD3 Ab (Ancell, Bayport, MN, USA) and incubated at 4°C overnight. The plates were washed 3 times with 10% FBS RPMI 1640 medium, and the CD4⁺ iNKT cells were added and cultured in a 5% CO₂ incubator at 37°C. After 7 d, the supernatant was collected and passed through a 0.22 µm filter to remove cellular debris.

Purification of DC maturation factors with FPLC

To purify DC maturation factors, we applied protein samples to an AKTA Explorer 100 FPLC instrument (GE Healthcare, Piscataway, NJ, USA). In brief, 50 ml of anti‐CD3 Ab‐activated supernatant from iNKT cells was loaded onto an anion exchange chromatography column (5 ml; HiTrap DEAE FF; GE Healthcare) and eluted with 20 mM Tris‐Cl (pH 8.0) containing 1 M NaCl. This step was repeated 6 times to load 300 ml of sample. To determine the effect of the factors on DC maturation, we removed 50 µl samples of supernatant from each collection tube. The fractions that exhibited DC maturation effects were freeze dried and dialyzed with 20 mM Tris‐Cl binding buffer at 4°C overnight. The active fractions were reloaded onto cation exchange chromatography columns (5 ml; HiTrap CM FF, GE Healthcare), and the active fractions were identified by following the procedure described above for anion exchange column chromatography. The active fractions were reloaded onto affinity chromatography columns (1 ml; HiTrap Blue HP; GE Healthcare) and eluted with 50 mM KH₂PO₄ containing 1.5 M KCl (pH 7.0). The protein content in the final active fractions was analyzed with mass spectrometry.

Mass spectrometry

The partially purified DC maturation factors from the CD4⁺ iNKT cell culture supernatant were analyzed with 12% SDS‐PAGE performed at 35 mA for 40 min. The gels were stained with Coomassie Brilliant Blue G‐250 (Boston BioProducts, Worcester, MA, USA), and the areas of interest were manually excised and digested in‐gel with trypsin. Samples were analyzed on an LTQ‐FT Plus Mass Spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Each reconstituted sample was separated on a C18 column (75 μm × 20 cm) at a flow rate of 375 μl/min with a gradient of 5–60% water/0.1% formic acid and acetonitrile/0.1% formic acid over 150 min. The instrument was operated in a top‐10 configuration with the full Fourier transform scan acquired at a resolution of 100 K with the MS2 selections performed in the ion trap.

Peptides were identified by using Sequest through the Bioworks Browser 3.2, with fragment data searched against the RefSeq Human Database from the National Center for Biotechnology Information (Bethesda, MD, USA) by using differential carbamidomethyl‐modified cysteines and oxidized methionines. Peptide score cutoff values were chosen at an Xcorr score of 1.8 for singly charged ions, 2.5 for doubly charged ions, and 3.0 for triply charged ions, along with deltaCN values of 0.1 and RSp values of 1. These parameters included less than 1% of the dummy reverse database. The cross‐correlation values chosen for each peptide ensured a high‐confidence match for the various charge states, and the deltaCN cutoff value ensured the uniqueness of the peptide hit. The RSp value of 1 ensured that the peptide matched the top hit in preliminary scoring and that the .dta peptide fragment file matched only 1 protein.

DC generation from buffy coats and the maturation test

Buffy coats were obtained from healthy donors according to the institutional guidelines of the University of Florida. PBMCs were prepared via density centrifugation by using Ficoll‐Paque gradients (Amersham Pharmacia Biotech, Uppsala, Sweden). Seventy million PBMCs were added to 10 cm tissue culture plates in 10 ml RPMI 1640 medium supplemented with 10% heat‐inactivated FBS, 2 mM l‐glutamine, 100 U/ml penicillin, and 10 µg/ml streptomycin and incubated for 1 h in a 5% CO₂ incubator at 37°C. Nonadherent cells were removed, and adherent cells were collected with a cell scraper. In some instances, PBMCs were mixed with microbead‐conjugated anti‐CD14 Ab (Miltenyi Biotech, Auburn, CA, USA) and incubated on ice for 30 min. After a wash with binding buffer [2% FBS, 2 mM EDTA, 20 mM PBS (pH 7.4)], the CD14⁺ cells were purified with autoMACS (Miltenyi Biotech). The purified cells were subsequently cultured in 10 cm culture plates in RPMI 1640 medium supplemented with 10% heat‐inactivated FBS, 10 mM HEPES, 2 mM l‐glutamine, 100 U/ml penicillin, 10 µg/ml streptomycin, 50 μM 2‐ME, 5 ng/ml GM‐CSF, and 10 ng/ml IL‐4 (R&D Systems, Minneapolis, MN, USA). Half of the medium was replaced on d 3 and 6.

For DC maturation analysis, the nonadherent cells (iDCs) were harvested on d 7 and stimulated with TNF‐α or purified supernatant from NKT cells. FACS with a FACSCalibur instrument (BD Bioscience) was used to determine the maturation of the iDCs by measuring the levels of CD86, PDL‐1, and HLA‐DR on the cell surface.

Flow cytometry

The phenotypes of iNKT cell clones were characterized by using a panel of cell surface Abs conjugated to fluorescein isothiocyanate or phycoerythrin. In brief, the cells were washed in staining buffer (PBS [pH 7.2] supplemented with 2% FBS and 0.1% sodium azide), incubated at 4°C for 30 min with the appropriate conjugated Abs, and washed twice with staining buffer. The fluorescein isothiocyanate–conjugated Abs were CD4, CD8, and IgG isotype. The phycoerythrin‐conjugated Abs were CD3, CD4, CD8, and NKT (6B11). CD14, HLA‐DR, PDL‐1, and CD86 were used to assess DC maturation. Intracellular staining was performed according to the manufacturer's recommendations (BD Pharmingen, San Diego, CA, USA). Brefeldin A solution (1 × 40 µl per well; BD Biosciences) was added during the last 6–12 h of culture. After staining for extracellular antigens, the cells were incubated with 500 µl of 1× permeabilizing solution 2 (BD Biosciences) for 10 min at room temperature in the dark. Isotype‐matched Abs were included for all experiments to control for nonspecific binding. Forward‐ and side‐scatter gating excluded dead cells. Flow cytometry was performed with a FACSCalibur (BD Biosciences) instrument, and the data were analyzed by using FlowJo software (TreeStar, Ashland, OR, USA).

Western blot analysis

CD4⁺ iNKT cells activated with anti‐CD3 Ab (Ancell Co., Bayport, MN, USA) were harvested and washed twice with ice‐cold PBS. The pellet was resuspended in RIPA buffer (50 mM Tris‐HCl [pH 7.4], 150 mM NaCl, 1% NP‐40, 0.5% sodium deoxycholate, 0.1% SDS, 5 mM EDTA, and 1 mM EGTA; Boston BioProducts) containing a proteinase inhibitor cocktail (Roche). After incubation on ice for 30 min, the lysate was centrifuged at 14,000 rpm for 10 min, and the supernatant was collected for analysis. One volume of Laemmli's sample buffer (4% SDS, 20% glycerol, 10% 2‐ME, and 4 mg/100 ml bromophenol blue) was added to the supernatant, which was then incubated at 95°C for 5 min.

The samples were analyzed with 12% SDS‐PAGE performed at 35 mA for 40 min. The supernatants from the anti‐CD3 Ab‐activated CD4⁺ iNKT cells were filtered with 0.2 μm membranes, mixed with Laemmli's sample buffer, and applied to the gels. The proteins were electrotransferred to PVDF membranes at 300 mA for 2 h with transfer buffer containing 25 mM Tris, 190 mM glycine, and 20% methanol. After 2 h of saturation at room temperature in PBS/0.05% Tween 20 containing 5% nonfat milk (Bio‐Rad, Hercules, CA, USA), the membranes were incubated with anti‐S100A8 Ab or anti‐β tubulin Ab (1:500 v/v; Cell Signaling Technology, Danvers, MA, USA) at 4°C overnight. The membranes were washed 3 times with PBS/0.05% Tween 20 and incubated with HRP‐conjugated anti‐mouse IgG (1:1000 v/v) and anti‐rabbit IgG Ab (1:1000 v/v; Cell Signaling Technology) for 1 h at room temperature. The proteins were detected with a Western Lightning Plus‐ECL kit (PerkinElmer, Waltham, MA, USA), according to the manufacturer's instructions.

Cytokine assay

The amounts of G‐CSF, GM‐CSF, IL‐10, IL‐12 (p70), IL‐1β, IL‐2, IL‐4, IL‐5, IL‐6, IL‐7, IL‐8, IL‐17, IFN‐γ, MCP‐1, MIP‐1β, and TNF‐α in the supernatants from the cultured cells were measured with a Bio‐Plex Pro human cytokine 17‐plex assay kit (Bio‐Rad) according to the manufacturer's instructions.

Multiplex transcription factor assay

To investigate the transcription factors of matured DCs differentiated by LPS and S100A8, we treated iDCs with LPS and S100A8 for 6 and 48 h. After harvesting the stimulated iDCs, we washed them 3 times with ice‐cold PBS. Nuclear extracts from the DCs were prepared with a Procarta Transcription Factor assay kit (Panomics, Fremont, CA, USA), according to the manufacturer's instructions. Each nuclear extract sample (0.1–10 μg/ml) was combined with 10 μl of a 24 biotin‐labeled cis‐element probe mixture that contained 2× binding buffer in a total reaction volume of 20 μl [18]. As a control, 10 μl distilled water was added to a sample of the reaction mixture instead of nuclear extract. The mixture was incubated for 30 min at 15°C by using a thermocycler. The transcription factor–probe complex was applied to the spin column included with the Protein/DNA Array kit (Panomics). The probes that formed complexes with active transcription factors were recovered from the spin column. Twenty microliters of the eluted sample was hybridized to the 24‐plex microspheres, and probe recovery was detected by using the Luminex 100 multiplexing system (Luminex, Austin, TX, USA). We evaluated the following transcription factors in mature DCs with LPS and S100A8: AP‐1, ATF‐2, CCAAT, CCAAT displacement protein, CCAAT enhancer‐binding protein, c‐MYC, cAMP, E2F transcription factor 1, early growth response, ELK‐1, ETS/PEA, GATA, HIF‐1, HSF‐1, IFN regulatory factor, IFN stimulatory element, NF of activated T‐cells, NF‐κB, NF‐Y, p53, peroxisome proliferator‐activated receptor, runt‐related transcription factor/acute myeloid leukemia, SMAD, STAT‐1, STAT‐3, transcription factor II D, and X‐box binding protein 1.

Statistical analyses

All values are expressed as means ± sd. To identify genes with statistically significant differences between T‐cell subsets or cell types, we performed 1‐way ANOVA with a P‐value cutoff of 0.01 and used the Benjamini and Hochberg false‐discovery rate for multiple testing corrections. Statistical significance in cluster analyses was determined with 1‐way ANOVA with a P‐value cutoff of 0.05 and the Benjamini and Hochberg false‐discovery rate. The results of other experiments, including the differential effect of CD4⁺ and DN iNKT cells on DC maturation, were analyzed with 1‐way ANOVA followed by Bonferroni's multiple‐comparison test. P < 0.05 indicated statistical significance.

RESULTS

CD4⁺ iNKT cells induce DC maturation

To assess the DC maturation capabilities of iNKT cells, we initially isolated iNKT cells from the PBMCs of healthy volunteers by using the 6B11 Ab, which reacts specifically with the TCR CDR3 region of human iNKT cells [15] and a microbead‐conjugated anti‐IgG1 Ab. Next, iNKT cells were separated into CD4⁺ and DN cells by using CD4 microbeads, which were evaluated with flow cytometry ( Fig. 1A ). The soluble factors of iNKT cells have been shown to have DC maturation capability [6, 19]. Therefore, supernatants were prepared from anti‐CD3 Ab‐activated CD4⁺ and DN iNKT cells. The cells were washed with complete culture medium 3 times, applied to plates coated with 100 ng/ml anti‐CD3 Ab and incubated with 50 U/ml IL‐2. The supernatant was collected from both CD4⁺ and DN iNKT cells and added to cultures of iDCs that were induced from freshly purified PBMCs via CD14 magnetic sorting and incubated with GM‐CSF and IL‐4. After 48 h, representative surface markers of mature DCs, HLA‐DR, and CD86, were measured by using flow cytometry [11]. These markers were significantly upregulated after the addition of supernatant from CD4⁺ iNKT cell clones (iNKT‐27 and ‐67) compared with DN iNKT cell clones (iNKT‐B5, ‐H9, and ‐A13; Fig. 1B, C). The increased number of cells that were positive for both CD86 and HLA‐DR had a FACS profile (Fig. 1B) representative of the CD4⁺ iNKT cell clone iNKT‐67, as shown in see Fig. 1C. Double‐positive CD86 and HLA‐DR cells increased from 52.1% in untreated control DCs to 81.1% in the supernatant of CD4⁺ iNKT cell‐treated DCs. These observations prompted us to focus on the soluble factors of CD4⁺ iNKT cells for further experimentation.

Figure 1.

DC maturation and regulatory T‐cell generation in response to CD4⁺ or DN iNKT cells. (A) FACS profile of representative CD4⁺ and DN iNKT cells. (B) Human iDCs were differentiated into mature DCs by adding supernatant from anti‐CD3 Ab‐activated CD4⁺ iNKT cells and DN iNKT cells. The data were obtained from triplicate experiments with each clone. (C) Representative FACS profile of DC maturation induced by PBS only (control) or the culture supernatant from CD4⁺ iNKT 67. (D) Compared with DN iNKT cells, CD4⁺ iNKT cells preferentially enhanced Foxp3 and CD25 expression. (E) iDCs were matured with supernatant from CD4⁺ iNKT cells in Transwells. After 4 d of stimulation with anti‐CD3 and ‐CD28 Abs, CFSE⁺CD4⁺T cells were analyzed for Foxp3 and CD25 expression. Significant differences were evaluated with 1‐way analysis of variance followed by Bonferroni's multiple comparison tests. ***P < 0.0001. The results represent 3 independent experiments. Error bars indicate the means ± sd.

CD4⁺ iNKT cells generate T_reg cells and inhibit T_H1 and T_H17 cell development

Evidence has suggested that iNKT cells could expand the CD4⁺CD25⁺Foxp3⁺ T_reg‐cell population [20]. However, the mechanism of iNKT cell function has not been well defined. iNKT cell subsets were tested for their differential capacity to convert CD4⁺CD25⁻Foxp3⁻ T cells into T_reg cells in a series of in vitro reconstitution experiments. The expression of CD25 and Foxp3 in a coculture of CD4⁺CD25⁻Foxp3⁻ T cells with DN iNKT cells was much lower than that with CD4⁺ iNKT cells (Fig. 1D). Because iNKT cells are believed to control myeloid APC effector function and thereby regulate T‐cell differentiation [21, 22], the coculture experiments in Transwell plates (Corning, Inc., Corning, NY, USA) were repeated with the addition of autologous CD14⁺‐derived APCs. The iNKT cell‐dependent conversion was markedly enhanced by the inclusion of myeloid APCs, and the conversion was more effective when CD4⁺ rather than DN iNKT cell lines were used.

In addition, the generation of T_reg cells by CD4⁺ iNKT cells was not caused by cell‐to‐cell contact (Fig. 1E). GalCer stimulation resulted in an expansion of T_reg cells from non‐iNKT (6B11⁻) CD25⁻CD4⁺ T cells ( Fig. 2A ). As observed for iNKT cell expansion, IL‐2 neutralization suppressed the expansion of T_reg cells in response to α‐GalCer, whereas IL‐15 neutralization did not. Therefore, these results indicate that the IL‐2 secreted by the α‐GalCer‐stimulated iNKT cells enhanced the expansion of T_reg cells. Strikingly, iNKT cell‐dependent conversion was markedly enhanced by the inclusion of myeloid APCs, and the conversion was 3‐ to 5‐fold more effective when CD4⁺ rather than DN iNKT cell lines were used. Owing to phenotypic similarity, T_reg cells converted by CD4⁺ iNKT cells were designated iNKT‐T_reg cells (Fig. 2B).

Figure 2.

CD4⁺ iNKT cells selectively generate functional T_reg cells (CD4⁺CD25⁺Foxp3⁺) from non‐T_reg cells (CD4⁺CD25⁻Foxp3⁻) in an IL‐2‐ and APC‐dependent manner. (A) CFSE‐labeled PBMCs were stimulated with α‐GalCer in the presence of the indicated neutralizing Abs for 5 d. CD4⁺ T cells (gated at 6B11⁻) were analyzed for the percentage of CD4⁺CD25⁺Foxp3⁺ cells and CFSE dilution. (B) Established iNKT‐T_reg cells were analyzed for purity and Foxp3, CD25, and glucocorticoid‐induced TNF receptor expression. (C) CFSE‐labeled CD4⁺CD25⁻ T cells were stimulated with anti‐CD3 and anti‐CD28 Abs plus CD14⁺ APCs in the absence (white bar) or presence (black bars) of iNKT‐T_reg cells or presence of CD4⁺CD25⁺ T cells (gray bars) for 4 d. The CFSE‐labeled T_reg or iNKT‐T_reg cells were also stimulated with anti‐CD3 and ‐CD28 Abs plus CD14⁺ APCs. Proliferation of CFSE⁺ cells was analyzed with the CFSE dilution method. The data are means ± sd of triplicate experiments.

The iNKT‐T_reg cells were analyzed for their suppressive activity on the proliferation of CD4⁺CD25⁻ T cells. Indeed, the iNKT‐T_reg cells and T_reg cells were equally efficient in suppressing the proliferation of CD4⁺CD25⁻ T cells (Fig. 2C). The generation of T_reg cells by CD4⁺ iNKT cells is intriguing, because it explains the immunoregulatory functions of iNKT cells in inflammatory disease. Therefore, it seems reasonable to postulate that CD4⁺ iNKT cells regulate the differentiation of proinflammatory T_H1 and T_H17 cells by generating T_reg cells ( Fig. 3A and B ). The inclusion of CD4⁺ iNKT cells or iNKT‐T_reg cells with CD4⁺CD25⁻Foxp3⁻ T cells and CD14⁺‐derived APCs significantly suppressed the differentiation of T_H1 and T_H17 cells (Fig. 3C–F). Even under conditions that efficiently induced T_H1 and T_H17 cell development, the inclusion of CD4⁺ iNKT cells suppressed this pathway and skewed the outcome toward the development of T_reg cells (Fig. 3F), whereas iNKT‐T_reg cells suppressed all forms of differentiation (Fig. 3C). These results allow us to hypothesize that T_reg cells suppress the differentiation of proinflammatory T_H1 and T_H17 cells, even under proinflammatory conditions.

Figure 3.

CD4⁺ iNKT cells suppress T_h1 and T_h17 cell differentiation by generating T_reg cells. CFSE‐labeled CD4⁺CD25⁻ T cells were stimulated with anti‐CD3 and ‐CD28 Abs plus CD14⁺ APCs (A) in the presence of CD4⁺ iNKT (B) or iNKT‐T_reg cells (C) for 5 d. CFSE⁺CD4⁺ T cells were analyzed for IL‐17⁺ (D), IFN‐γ⁺ (E), and Foxp3⁺ (F) cells. The graphs show present the means ± sd of triplicate experiments. The results represent 3 experiments.

Neutralizing Abs against several cytokines did not block DC maturation

Human and murine iNKT cells have been implicated in the differentiation of tolerogenic myeloid DC subsets [11, 20], and the CD4⁺ iNKT cell phenotype appears to correspond to an anti‐inflammatory profile [19, 23]. The supernatant from activated CD4⁺ iNKT cells induces myeloid DC differentiation [11]. In the characterization of CD4⁺ and DN iNKT cells, it was important to ascertain the cytokines that these cell types secrete after activation. We used the same pair of iNKT cell lines, CD4⁺ and DN, and harvested the supernatants 24 h after anti‐CD3 Ab activation. Assay of the supernatants on a RayBio human cytokine Ab array of ∼120 different cytokines showed that the inflammatory mediator ICAM [24, 25] and TIMP‐1 [26], as well as the inflammatory cytokines IL‐5, IL‐6, IL‐8, IL‐13, and GM‐CSF, were detected in the supernatants of both types of iNKT cells (Table 1). Of these molecules, TIMP‐1 and IL‐8 have not been described as products of iNKT cells.

Table 1.

Cytokine and chemokine profiling of supernatant of CD4⁺ and DN iNKT cells

Intensity	CD4+ iNKT	DN iNKT
High a	IL‐5, IL‐8, IL‐13, GM‐CSF, TNF‐α, TNF‐β	IL‐5, IL‐6, IL‐8, IL‐13, GM‐CSF, IFN‐γ, TNF‐α, TNF‐β
Medium b	IL‐6, MIP‐1 β	TIMP1, MIP‐1 β
Low c	IL‐2, Acrp30, ICAM‐1, Oncostatin M, RANTES, sTNFRII, TIMP1, uPAR	Acrp30, FGF‐4, ICAM‐1, NAP‐2, Oncostatin M, RANTES, sTNFRII, uPAR

Acrp = adipocyte complement‐related protein, RANTES = regulated on activation, normal T cell expressed and secreted, uPAR = urokinase‐type plasminogen activator

^aCytokines detected within a minute in the ELISA reaction.

^bCytokines detected between 1 and 5 min in the ELISA reaction.

^cCytokines detected between 5 and 10 min in the ELISA reaction.

However, the iNKT cell types differed in that the DN iNKT cells produced IFN‐γ but the CD4⁺ iNKT cells did not. The panel of cytokines and chemokines secreted by both the CD4⁺ and DN iNKT cells could be expected to induce DC maturation, but only TNF‐α is a strong stimulus for the terminal differentiation of myeloid DCs. Most of the cytokines and chemokines derived from iNKT cells (i.e., GM‐CSF, IL‐4, MIP‐1α, and MIP‐1β) are more crucial for the early maturation of myeloid DCs. IL‐10, IL‐13, and IFN‐γ have been shown to regulate effector function instead of inducing terminal differentiation.

Therefore, we examined the supernatants of anti‐CD3 Ab‐stimulated CD4⁺ and DN iNKT cells and were unable to discern any significant differences in the levels of these cytokines secreted into the media. The addition of anti‐IL‐4, anti‐IL‐10, anti‐IL‐13, IFN‐γ, and TNF‐α mAbs did not completely inhibit the DC maturation effect of the supernatant prepared from the CD4⁺ iNKT cells (Supplemental Fig. 1A). In addition, all combinations of IL‐4, IL‐13, and GM‐CSF lacked the DC maturation effect (Supplemental Fig. 1B), and IL‐13 did not induce T_reg‐cell generation (Supplemental Fig. 1C). We found no reasonable candidates for promoting DC maturation. Therefore, we investigated DC maturation factors originating from the supernatant of CD4⁺ iNKT cells by using mass spectrometry.

Purification of DC maturation factors from CD4⁺ iNKT cells

Previous data have shown that DC maturation factors do not belong to the cytokine or chemokine families of proteins. In addition, the DC maturation effect is not due to cell‐to‐cell contact. Therefore, we prepared culture supernatant of CD4⁺ iNKT cells activated by anti‐CD3 Ab to purify the DC maturation factors. The prepared supernatant was applied to anion exchange, cation exchange, and affinity chromatography columns as described in Materials and Methods. We then investigated the active regions of anion exchange column chromatography (unbound fractions, collection tubes 3–15; Fig. 4A ), cation exchange column chromatography (bound fractions, collection tubes 27–42; Fig. 4B), and affinity chromatography (bound fractions, collection tubes 9–17; Fig. 4C). DC maturation states were examined by assaying cell surface expression levels of HLA‐DR and CD86 with FACS. We designated as active fractions the samples containing DCs that exhibited high surface expression levels of HLA‐DR and CD86. The final active regions of affinity chromatography were analyzed with SDS‐PAGE, after which we used mass spectrometry for protein identification.

Figure 4.

Purification of DC maturation factors secreted by CD4⁺ iNKT cells with FPLC and the expression of S100A8 in CD4⁺ iNKT cells. (A) Anti‐CD3 Ab‐activated supernatants from CD4⁺ iNKT cells were loaded onto an anion exchange chromatography column and eluted with 20 mM Tris‐Cl (pH 8.0) containing 1 M NaCl. To determine the effect on DC maturation, 50 µl of supernatant was taken from each collection tube. The collection tubes in which factors showed effects on DC maturation were freeze dried and dialyzed with 20 mM Tris‐Cl binding buffer at 4°C overnight. (B) The active region was reloaded onto a cation exchange chromatography column, and the active fractions were identified as described for DEAE FF. (C) The active fraction prepared using CM FF was reloaded onto an affinity chromatography column and eluted with 50 mM KH₂PO₄ containing 1.5 M KCl (pH 7.0). The protein content of the final active fraction was analyzed with mass spectrometry. ○, protein concentration; ▪, DC maturation capacity as assessed via FACS analysis with HLA‐DR and CD86 Abs. RT‐PCR (D), Western blot analysis (E), relative expression and secretion levels of S100A8 (F) , and relative expression level of S100A8 in cells (G), according to the concentration of anti‐CD3 Ab.

Effect of S100A8 on DC maturation

Several noteworthy proteins were obtained in the active chromatography fraction that demonstrated effects on DC maturation in 9 different experiments. The candidate proteins were sorted according to the number of hits from 9 independent purification series (Table 2). Among them, S100A8 was detected 8 times in 9 analyses, and PIP was detected 7 times. ATP‐binding cassette, desmoplakin isoform 1, haptoglobin, serpin peptidase inhibitor clade A (member 3), transferrin, and transthyretin were detected 6 times, whereas ceruloplasmin, hemopexin, mesotrypsin, orosomucoid 1, and serine (or cysteine) proteinase (clade B) were detected 5 times. S100A9, the counterpartner of S100A8, was detected 3 times, along with arginase I. Among these components, we chose several candidates and tested their effects on DC maturation. These candidates included arginase I (R&D Systems), which was identified only 3 times, but was a possible hit; mesotrypsin (rhTrypsin‐3, R&D Systems); S100A8 (Sino Biologic, Beijing, China); S100A9 (Sino Biologic); and PIP (OriGene). The results of Western blot analysis verified that these proteins were expressed in CD4⁺ iNKT cells. Among them, S100A8 and PIP demonstrated an effect on DC maturation. We focused on the characteristics of S100A8 in this study.

Table 2.

The supernatant of CD4⁺ iNKT cells’ protein components analyzed by mass spectrometry

Group a	Protein
A	S100A8
B	PIP
C	ATP‐binding cassette
	Desmoplakin isoform 1
	Haptoglobin
	Serpin peptidase inhibitor clade A, member 3
	Transferrin
	Transthyretin
D	Ceruloplasmin
	Hemopexin
	Mesotrypsin
	Orosomucoid 1
	Serine (or cysteine) proteinase inhibitor, clade B

^aA: proteins detected 8 times of 9 sample analysis of 4 independent experiments. B: proteins detected 7 times. C: proteins that were detected 6 times. D: proteins detected 5 times.

The gene expression of S100A8 was confirmed in CD4⁺ iNKT cells with RT‐PCR (Fig. 4D). This expression was also verified in cells, and we used Western blot analysis to check the existence of S100A8 protein in the culture supernatant of CD4⁺ iNKT cells (Fig. 4E–G). In the supernatant, the level of S100A8 expression increased as the concentration of anti‐CD3 Ab increased (Fig. 4E, F). Next, we tested whether anti‐S100A8 Ab blocked the DC maturation induced by culture supernatant of CD4⁺ iNKT cells by treating iDCs with 1 µg/ml anti‐S100A8 Ab (R&D Systems) or IgG (control) for 1 h before adding the supernatant from CD4⁺ iNKT cells. The maturation of iDCs was assessed by measuring the DC surface expression levels of CD86 and HLA‐DR with FACS. The anti‐S100A8 Ab signal decreased to 34.2% in the iDC maturation‐induced culture supernatant of CD4⁺ iNKT cells (P = 0.0001; Fig. 5A ).

Figure 5.

Blocking effect of anti‐S100A8 and anti‐TLR‐4 on DC maturation and inhibition of T_reg generation by S100A8 siRNA treatment in CD4⁺ iNKT cells. (A) A concentration of 1 µg/ml anti‐S100A8 Ab or IgG control was preincubated with the supernatant from CD4⁺ iNKT cells before iDCs were added. iDC maturation was assessed by measuring the DC surface expression levels of CD86 and HLA‐DR as determined with FACS (t test, P = 0.0001). (B) Effects of recombinant S100A8 and S100A9 and the S100A8‐S100A9 complex on DC maturation (n = 2). (C) Comparison of the effects of S100A8 and TNF‐α on DC maturation. iDCs were treated with PBS as a control, 1 µg/ml S100A8, 1 µg/ml S100A8, and 1 µg/ml anti‐S100A8 Ab, and 100 ng/ml TNF‐α. (D) The data indicate the relative ratios of the various treatments. (E) Maturation effects of various concentrations of S100A8 on iDCs. PDL‐1 expression in DCs after stimulation with LPS, Nɛ‐carboxymethyllysine, and S100A8 with or without anti‐TLR‐4 (F) or anti‐RAGE Ab (G). (H) Reduced T_reg‐cell generation after treatment of CD4⁺ iNKT cells with S100A8 siRNA (10 nM). (I) Comparison of T_reg cell generation by S100A8 and LPS in PBMCs. *P < 0.05; **P < 0.001; ***P < 0.0001. Results represent 3 independent experiments. Error bars indicate means ± sd. See also Supplemental Fig. 3.

When we compared this result to the effects of the individual subunits of the S100A8‐S100A9 complex on iDC maturation, we found that both S100A8 alone and the S100A8‐S100A9 complex induced iDC maturation. S100A9 did not induce the level of iDC maturation observed in response to S100A8 alone (P = 0.0386; Fig. 5B). We then compared the effect of recombinant S100A8 protein on iDC maturation to that of TNF‐α and found that S100A8 and TNF‐α had similar effects. Moreover, anti‐S100A8 Ab blocked the effect of S100A8 on DC maturation (Fig. 5C, D). The effect of S100A8 on DC maturation was concentration dependent (Fig. 5E).

S100A8 generates T_reg cells from naïve T cells through TLR‐4

S100A8 is known to bind RAGE as well as TLR‐4 and can be associated with inflammation [14, 27, 28–29]. DC maturation induced by S100A8 was blocked by the addition of an anti‐TLR‐4 Ab (Fig. 5F). Nɛ‐carboxymethyllysine, an agonist of RAGE, neither induced DC maturation nor inhibited the maturation‐inducing effect of S100A8 (Fig. 5G). CD4⁺ iNKT cells were treated with siRNA toward S100A8 to silence the S100A8 gene before adding it to immature DCs, which, compared with the untreated group, induced fewer T_reg cells (Fig. 5H). In addition, when S100A8 was added to PBMCs, T_reg cells were generated in an S100A8‐concentration–dependent manner (Fig. 5I). These observations may indicate that S100A8 signals for DC maturation are transduced via TLR‐4 signaling, similar to the case with LPS.

TLR activation with LPS is reportedly an on/off signal that controls proinflammatory or tolerogenic DC maturation after interaction with iNKT cells [30]. In addition, iNKT cells have long been known to have quixotic Janus‐like behavior with respect to immune regulation [22]. Given that both human and murine iNKT cells have been implicated in the differentiation of “tolerogenic” myeloid DC subsets [6, 21], we attempted to determine the combination of S100A8, anti‐CD1d, and MHC class II proteins needed to induce tolerogenic DCs. The DC surface expression levels of CD86 and PDL‐1 induced in iDCs by the combination of anti‐CD1d/MHC class II Ab and S100A8 were higher than those induced by any other combination or single agent, but HLA‐DR expression was unaffected ( Fig. 6A ). Several cytokines were secreted by S100A8‐treated DCs, including G‐CSF, GM‐CSF, IFN‐γ, IL‐10, IL‐6, IL‐8, MCP‐1, MIP‐1β, and TNF‐α (Fig. 6B). However, the combination of anti‐CD1d and anti‐MHC class II Abs clearly induced IL‐8. With the exception of IL‐8, the levels of most cytokines—particularly IL‐6—induced by S100A8 were reduced or blocked by combined treatment with anti‐CD1d and anti‐MHC class II Abs (Fig. 6B; Supplemental Fig. 2). Notably, we found that S100A8 activates DCs via TLR‐4 and unidentified receptors. Compared with LPS, S100A8 induces IL‐10, G‐CSF, and PDL‐1, but the induction of IL‐1β, IL‐12, TNF‐α, and IL‐6 is significantly weaker (Fig. 6B, C). The combination of anti‐CD1d and anti‐MHC class II Abs did not induce IL‐10 to the extent observed with S100A8 (Supplemental Fig. 2).

Figure 6.

DC maturation with a combination of S100A8, anti‐CD1d Ab, and anti‐MHC class II Ab. Stimulation with a combination of plate‐bound anti‐CD1d/MHC class II Abs iDC maturation (A), cytokine production (B), and lower production of proinflammatory cytokines in S100A8‐treated iDCs (C). iDCs were transferred onto anti‐CD1d/MHC class II Ab‐coated plates and cultured for 48 h with or without soluble S100A8 to induce tolerogenic DCs. Cytokines were analyzed with a human cytokine 17‐plex assay kit. *P < 0.05; **P < 0.001. Error bars indicate means ± sd. See also Supplemental Fig. 2.

Next, we evaluated whether S100A8‐induced mature DCs or a combination of anti‐CD1d and MHC class II Abs could induce T_reg cells. The S100A8‐treated DC group ( Fig. 7A, B ), the combination of anti‐CD1d and anti‐MHC class II Abs and S100A8‐treated DC groups treated with the combination of anti‐CD1d and anti‐MHC class II Abs generated T_reg cells, but T_reg cells were not generated in the anti‐CD1d and anti‐MHC class II Ab‐treated groups (Fig. 7A, B). In addition, T_reg‐cell generation was enhanced depending on the concentration of S100A8 (from 100 ng to 2 μg; Supplemental Fig. 3A) and reduced by S100A8 siRNA (Supplemental Fig. 3B).

Figure 7.

S100A8 and the combination of plate‐bound anti‐CD1d/MHC class II Abs stimulate iDCs to induce CD4⁺CD25⁺Foxp3⁺ T cells. (A) Representative FACS profiles of T_reg‐cell generation induced by the combination of S100A8, anti‐CD1d, and anti‐MHC class II Abs. (B) T_reg‐cell generation by the combination of S100A8, anti‐CD1d, and ‐MHC class II Abs. iDCs were transferred to anti‐CD1d/MHC class II Ab–coated plates and cultured for 48 h with or without soluble S100A8 to induce tolerogenic DCs. Tolerogenic DCs were cocultured with CFSE‐labeled CD4⁺CD25⁻Foxp3⁻ T cells with 1 µg/ml anti‐CD3, 500 ng/ml anti‐CD28, and 50 U/ml IL‐2 for 5 d. The development of T_reg cells was assessed by determining the expression levels of CD4, CD25, and Foxp3 via FACS analysis. See also Supplemental Fig. 3.

Transcription factors induced by S100A8 in tolerogenic DCs

We investigated the transcription factors potentially induced by S100A8 by activating iDCs with S100A8 and including an LPS‐treated iDC group as a positive control and an untreated iDC group as a negative control. Before preparing the samples for transcription factor assays, we used FACS to analyze the iDC maturation state after 6 h (Supplemental Fig. 4) and 48 h ( Fig. 8 ) of S100A8 and LPS treatment. We then analyzed the expression levels of 28 transcription factors with a Procarta Transcription Factor Plex kit (Affymetrix, Santa Clara, CA, USA). After 6 h of treatment, we found no significant increase in DC maturation state, and no changes were observed in the levels of several transcription factors, including NF‐κB, AP‐1, ATF‐2, and HSF‐1 (Supplemental Fig. 4). However, after 48 h of treatment, the expression of CD86, PDL‐1, and HLA‐DR significantly increased. In addition, we found that NF‐κB and AP‐1 were upregulated by both S100A8 and LPS treatment, and ATF‐2 and HSF‐1 were upregulated by S100A8 (Fig. 8).

Figure 8.

Profiling S100A8‐induced activation of transcription factors with a microsphere‐based active transcription factor assay system. Nuclear extracts were prepared from DCs treated with S100A8 and LPS for 48 h. The maturation states of iDCs were evaluated with FACS analysis for PDL‐1 and HLA‐DR expression levels (A, B). Nuclear extracts (2 μg each) were incubated with a biotin‐labeled probe mix. The protein‐bound probes were isolated, hybridized to the transcription factor probe‐conjugated microsphere mix, and analyzed. (C) LPS and S100A8 treatment significantly increased transcription factor levels. See also Supplemental Fig. 4.

DISCUSSION

DCs are key mediators of adaptive immunity. In the absence of bodily infection, DCs in peripheral tissue rest in an immature state with limited capacity to stimulate naïve T cells. However, once infection occurs, DCs mature in a process characterized by phenotypic changes that results in an improved potential to promote T‐cell responses. DC maturation can be induced directly via stimulation through DC‐expressed TLRs or indirectly via exposure to cytokines released by local immune or nonimmune cells stimulated through their own TLRs. iNKT cells promote enhanced T‐cell responses when activated by a powerful stimulus, such as the synthetic glycolipid α‐GalCer, which implies that iNKT cells can provide all of the signals required for DC activation [31].The mechanism controlling the generation of proinflammatory DCs, as opposed to tolerogenic DCs, remains a critical question in the ontogeny of autoimmunity.

Steinman and Nussenzweig [32] proposed that in the absence of infection or inflammation, a steady‐state flux of immature DCs captures and processes endogenous antigens [32]. These DCs then define immunologic “self” and induce tolerance by specifically silencing autoreactive T‐cells and promoting the development of T_reg cells [33]. Although iDCs appear to harbor the capacity to induce tolerance to self and foreign antigens, there are many recent examples of DCs with potent tolerogenic capacity differentiated both in vivo and in vitro [11, 34]. Notably, the Bluestone group [35] recently demonstrated that T_reg cells interact directly with DCs bearing islet antigens in vivo and that this interaction precedes the inhibition of T_h cell activation by DCs. The explanation for these apparently contradictory functions appears to lie in the remarkable capability of iDCs to differentiate into specific functional subtypes during maturation [36]. Depending on the lineage (i.e., myeloid or lymphoid) and maturation stimulus, DCs potently control T‐cell effector function and cytokine profiles [20. 37].

The exchange of information between DCs and T cells is not unilateral. Although DCs are necessary for the efficient priming of antigen‐specific lymphocyte responses, T cells are also essential for optimal DC maturation [37]. Several studies, including our own, have underscored the regulatory importance of the cross talk between DCs and iNKT cells in shaping immune response and controlling the relative percentage of DC subsets [11, 38].

Autoreactive iNKT cells have also recently been demonstrated to induce DC maturation directly and potently. Moreover, mature DCs exposed to iNKT cells produce more IL‐10 than IL‐12, a phenotype consistent with tolerogenic function [19]. When we examined the supernatants of anti‐CD3 Ab‐stimulated CD4⁺ and DN iNKT cells, we found that the CD4⁺ iNKT cells produced more T_reg cells by generating tolerogenic DCs. Conventional CD4⁺ T cells did not behave similarly (Supplemental Fig. S5). In addition, our results demonstrated that stimulation with anti‐CD3 Ab increased only the level of S100A8 in the supernatant. The levels of S100A8 remaining in the cells were similar regardless of the amount of anti‐CD3 Ab, indicating that the increase in S100A8 observed under CD3 stimulation results from secretion, not from cell death (Fig. 4). Therefore, we attempted to identify the factors that could induce tolerogenic DCs to generate T_reg cells. The addition of anti‐IL‐4, IL‐10, GM‐CSF, IFN‐γ, and TGF‐β1 mAbs did not inhibit the activity of the supernatant or the partially purified proteins from CD4⁺ iNKT cells (Supplemental Fig. 1A). Furthermore, combinations of mAbs did not abrogate the effect on DC maturation. The lack of inhibition by mAbs is consistent with cytokine array data and enzyme‐linked immunosorbent assay findings, which demonstrated no differences in the upregulation or secretion of these cytokines between CD4⁺ and DN iNKT cells ( Table 1 ; Supplemental Fig. 1A, B).

Several proteins were identified by mass spectrometry analysis ( Table 2 ). PIP was detected 7 times in this analysis. PIP binds to CD4 [39] and exerts a potent inhibitory effect on T‐lymphocyte apoptosis mediated by CD4 and TCR activation [40]. It also initiates fibronectin‐specific aspartyl protease activity [41]. PIP overexpression has been demonstrated in primary and metastatic breast cancers [42] and in some breast carcinoma cell lines. The exact function of this protein in mammary tumor progression remains unclear [43]. However, we discovered that PIP induced T_reg cells through tolerogenic DC generation (data not shown). Among them, S100A8 manifested the most hits. This molecule was expressed and secreted by iNKT cells (Fig. 4) and exhibited an effect on DC maturation (Fig. 5E) equal to those of LPS and TNF‐α (Fig. 5C, D). S100A8 forms a complex with S100A9 and is upregulated in inflammatory diseases such as sepsis, rheumatoid arthritis, inflammatory bowel disease, allograft rejection, lung disease, vasculitis, and cancer [44, 45, 46–47].

The S100A8‐S100A9 complex is secreted by activated phagocytes and has proinflammatory effects. The abnormally elevated expression of these proteins was recently identified in an inflammatory disorder characterized by recurrent infections and systematic inflammation, which also indicates a central role for the S100A8‐S100A9 complex in the inflammatory process. In addition, the extracellular function of these proteins was recently observed in sepsis. S100A8 is an active component that induces the intracellular translocation of myeloid differentiation primary response protein 88 and the activation of IL‐1 receptor‐associated kinase‐1 and NF‐κB, which results in the elevated expression of TNF‐α. S100A8 interacts with the TLR‐4‐MD2 complex, thus representing an endogenous ligand for TLR‐4. Therefore, the S100A8‐S100A9 complex is a novel inflammatory component that amplifies phagocyte activation upstream of TNF‐α‐dependent effects during sepsis [14].

Similar results have been reported by Petersen et al. [48], who demonstrated the upregulation of CD86 and MHC class II proteins on the surface of BMDCs in C57BL/6 mice after stimulation with S100A8 for 48 h and proliferation of CD4 and CD8 T cells with mature BMDCs. They described the role of S100A8 as proinflammatory in innate immunity and regulatory in adaptive immunity, which may be a key mechanism for preventing tissue damage owing to overwhelming immune responses [48].

During DC maturation, S100A9 did not enhance the function of S100A8 (Fig. 5B), and S100A8‐mediated signaling was transduced via TLR‐4 (Fig. 5F). A similar result reported previously showed that BMDC activation by S100A8 in mice was mediated by TLR‐4, whereas RAGE was not involved [48]. In addition, S100A8 generated T_reg cells in a concentration‐dependent manner (Supplemental Fig. 3), which was enhanced by the addition of anti‐CD1d and anti‐MHC class II Abs (Fig. 7). However, S100A8 was expressed in both CD4⁺ iNKT and DN iNKT cells. We hypothesize that CD4⁺ iNKT cells generate tolerogenic DCs that support the differentiation of T_reg cells from naïve T cells, and this result suggests that other candidates described above for DC maturation are expressed only in CD4⁺ iNKT cells or that other control mechanisms reduce or inhibit the function of S100A8 in DN iNKT cells.

Notably, 2 transcription factors, ATF‐2 and HSF‐1, were elevated by S100A8 stimulation in mature DCs, but were not elevated by LPS stimulation (Fig. 8). Various forms of cellular stress, including genotoxic agents, inflammatory cytokines, and ultraviolet irradiation, stimulate the transcriptional activity of ATF‐2 [49, 50]. ATF‐2 is activated by TGF‐β signaling and cooperates with Smad3 to mediate TGF‐β effects on chondrocyte maturation [51]. The relationship between TLR signaling and ATF‐2 expression was recently explained by Altmayr et al. [52], who showed that ATF‐2 is involved in the transcriptional regulation of TLR‐stimulated (especially TLR2) TNF‐α production in DCs [52]. However, whether ATF‐2 correlates with TLR‐4 signaling remains unclear.

In addition, ATF‐2 plays a central role in TGF‐β signaling by acting as a common nuclear target of both the Smad and TGF‐β activated kinase‐1 pathways [53]. The regulation of TNF‐α expression is rigorous and redundant. At the transcriptional level, LPS‐induced TNF‐α expression is regulated by NF‐κB, which interacts with multiple sites in the TNF‐α promoter [54] and the assembly of the core enhanceosome complex containing Sp1, ATF‐2, and Ets proteins on the proximal TNF‐α promoter [55, 56]. In this regulation mechanism, another transcription factor induced by S100A8‐stimulated DCs, HSF‐1 (Fig. 8), represses LPS‐induced TNF‐α transcription by binding to a heat‐shock response element in the murine TNF‐α 5′‐flanking sequence [57, 58–59]. In Caenorhabditis elegans, HSF‐1 interconnects the insulin/insulin‐like growth factor‐1, TGF‐β, and cGMP neuroendocrine systems to control development and longevity in response to diverse environmental stimuli. Furthermore, HSF‐1 upregulates another TGF‐β pathway‐interacting gene, daf‐9/cytochrome P450, thereby fine tuning the decision between normal growth and dauer formation [60].

The presence of endogenous Hsp70 in mice, guaranteed by the presence of its transcription factor HSF‐1, protects against induced colitis [61]. More recently, treatment with whole endotoxin‐free Mycobacterium tuberculosis‐Hsp70 inhibited the acute rejection of skin and tumor allografts [62]. Consequently, disease‐suppressive effects have been observed for both microbial and self (mammalian) Hsp70 in some studies of whole proteins and in others examining only peptides. The transfer of Hsp70 peptide‐induced T_reg cells suppresses ongoing experimental arthritis [63]. S100A8 appears to belong to damage‐associated molecular patterns such as Hsp70, which is expressed in stressed cells. It may share a signaling pathway or the transcription factor HSF‐1 to generate T_reg cells for tolerance to cell stress [64].

In conclusion, this study demonstrates that S100A8 is a factor secreted by CD4⁺ iNKT cells that induces tolerogenic DC maturation involved in T_reg generation ( Fig. 9 ). However, many other factors in the culture supernatant produced by iNKT cells remain to be elucidated, because a combination of S100A8 siRNA and PIP siRNA reduced T_reg generation compared to that resulting from the application of each siRNA alone (Fig. 5H; Supplemental Fig. 3B).

Figure 9.

Dendritic cell maturation induced by S100A8 secreted from CD4⁺iNKT cells. Anti‐CD3 Ab with anti‐CD28 Ab or α‐GalCer‐activated CD4⁺ iNKT cells secrete S100A8 or PIP, or both, to upregulate PDL‐1, MHC class II, and CD86 on the surface of DCs. Mature DCs also secrete IL‐10. Finally, both results lead to the induction of T_reg‐cell proliferation from naïve T cells.

AUTHORSHIP

H.W.L. conceived the study, performed the experiments, analyzed the data, prepared the figures, and wrote the manuscript. H.B.J. performed the T_h1 and T_h17 cell differentiation experiments. P.L.B. performed the cytokine profiling of supernatant of NKT cells. D.S. performed mass spectrometry. T.S.K. supervised the study. B.S.W. designed and supervised the study.

DISCLOSURES

The authors declare no conflicts of interest.

Supporting information

Supplementary Material Files

Supplementary Material Files