Natural killer gene complex–encoded immunomodulatory C-type lectin-like receptors include members of the NKRP1 and C-type lectin-like 2 (CLEC2) gene families, which constitute genetically linked receptor-ligand pairs and are thought to allow for NK cell–mediated immunosurveillance of stressed or infected tissues. The mouse C-type lectin-like receptor Nkrp1g was previously shown to form several receptor-ligand pairs with the CLEC2 proteins Clr-d, Clr-f, and Clr-g, respectively. However, the physiological expression of Nkrp1g and its CLEC2 ligands as well as their functional relevance remained poorly understood. Recently, we demonstrated a gut-restricted expression of Clr-f on intestinal epithelial cells that is spatially matched by Nkrp1g on subsets of intraepithelial lymphocytes. In this study, we investigated expression and ligand interaction of Nkrp1g in the splenic compartment, and found an exclusive expression on a small subset of NK cells that upregulates Nkrp1g after cytokine exposure. Whereas transcripts of Clr-d and Clr-f are virtually absent from the spleen, Clr-g transcripts were abundantly detected throughout different leukocyte populations and hematopoietic cell lines. However, a newly generated anti–Clr-g mAb detected only residual Clr-g surface expression on splenic monocytes, whereas many hematopoietic cell lines brightly display Clr-g. Clr-g surface expression was strongly upregulated on splenic CD8α+ conventional dendritic cells (DCs) and plasmacytoid DCs upon TLR-mediated activation and detectable by Nkrp1g, which dampens NK cell effector functions upon Clr-g engagement. Hence, different to the intestinal tract, in the spleen, Nkrp1g is selectively expressed by a subset of NK cells, thereby potentially allowing for an inhibitory engagement with Clr-g-expressing activated DCs during immune responses.
Natural killer cells are cytotoxic innate lymphocytes that eliminate transformed or infected cells upon recognition through a broad variety of germline-encoded inhibitory and activating receptors, but also secrete cytokines such as IFN-γ (1–3). A major group of NK cell receptors are C-type lectin-like receptors encoded by the NK gene complex (NKC) in both man and mouse (4–7). These include NKG2D and CD94/NKG2x receptors, but also various members of the NKRP1 and C-type lectin-like 2 (CLEC2) families that have been shown to establish genetically linked receptor-ligand pairs (8–10). Whereas NKRP1 family members are expressed by NK cells and T cells, the expression pattern of their CLEC2 ligands is more diverse (10, 11). In humans, these NKC-encoded C-type lectin-like receptor-ligand pairs include activation-induced C-type lectin (AICL, CLEC2B) and the activating NK receptor NKp80 (12, 13), keratinocyte-associated C-type lectin (KACL, CLEC2A), and the activating receptor NKp65 (14, 15), as well as lectin-like transcript 1 (LLT1, CLEC2D) engaging the inhibitory receptor NKR-P1A/CD161 (16, 17). In mice, nonhomologous but structurally related members of the NKRP1 and CLEC2 gene families also constitute several receptor-ligand pairs (8, 18). Of these, the inhibitory receptor Nkrp1b and its ligand Clr-b have been studied most extensively. Clr-b is broadly expressed by hematopoietic as well as nonhematopoietic tissues and its expression is downregulated on infected, stressed, or malignant cells (19–21). Reduced Clr-b surface expression renders such cells more susceptible to NK cell cytolysis due to diminished inhibitory signals provided by Nkrp1b (19, 22). Hence, this receptor-ligand pair has then proposed to represent another MHC class I–independent missing-self recognition system (19, 23–26). This hypothesis was recently supported by the finding that the murine CMV (MCMV) glycoprotein m12 acts as a Clr-b substitute assuring inhibitory Nkrp1b ligation and viral immunoevasion from NK cell recognition even when Clr-b is downregulated in the course of MCMV infection (27). Of note, the viral immunoevasin m12 also binds to the activating receptors Nkpr1a and Nkrp1c that have remained orphan receptors (27). In contrast, several CLEC2 family members have been reported as ligands for the related Nkrp1f and Nkrp1g receptors: both Nkrp1 receptors are capable of engaging Clr-d and Clr-g, whereas Nkrp1f binds in addition to Clr-c, and Nkrp1g to Clr-f (8, 18, 28, 29). With regard to Clr-f, we recently reported a pronounced tissue-specific expression on intestinal epithelial cells (IEC) that is spatially matched by the selective intestinal expression of Nkrp1g on a subset of intestinal intraepithelial lymphocytes (IEL) facilitating immunosurveillance of the gut epithelium by certain γδ IEL (30). Similarly, we recently described an exclusive expression of the orphan CLEC2 molecule Clr-a on mouse gut epithelium, supporting the idea that CLEC2 family members may fulfill tissue-specific immune functions (31, 32). Tissue-specific expression may also provide a rationale for the redundancy of Nkrp1g ligands. Whereas little is known about the expression of Clr-d, several studies detected Clr-g transcripts in hematopoietic cells, including lymphokine-activated killer (LAK) cells (5), bone marrow–derived dendritic cells (DCs), and macrophages (8), B cells, and T cells (33), as well as preferentially in hematopoietic organs (30, 31).
NK cells are known to sculpt immune responses by interacting with DCs (34–41). In mice, there are three splenic DC populations, namely CD8α+ and CD8α− conventional DCs (cDCs) and plasmacytoid DCs (pDCs) (42, 43). These subtypes exert different functions and express nonredundant pattern recognition receptors, allowing them to specifically react to insults. CD8α+ cDCs are the main population expressing TLR3, thereby sensing dsRNA, CD8α+ and CD8α− cDCs equally express TLR4 recognizing LPS, and pDCs mainly express TLR7 and 9, which sense ssRNA or CpG DNA, respectively (44–46). TLR-activated DCs can shape innate immune responses through activation of NK cells. Conversely, the NK cells can also regulate adaptive immune responses through impacting on maturation and activation of DCs, both by cell-cell contact or soluble factors (47).
In this study, we report the exclusive expression of the inhibitory immunoreceptor Nkrp1g in the spleen on a minor subset of mouse NK cells that is spatially matched by expression of the Nkrp1g-ligand Clr-g on activated splenic DCs. Therefore, Nkrp1g on activated NK cells may dampen NK responses toward splenic DCs in the course of immune responses.
Materials and Methods
BALB/c, C57BL/6, C3H, 129, and CD-1 ISG mice were purchased from Envigo (Horst, the Netherlands) or Charles River Laboratories (Sulzfeld, Germany). LOU/C rats were from Envigo. NOD SCID (NOD.Cg-Prkdcscid/J) and NOD SCID GAMMA mice (NOD.Cg-Prkdcscid Il2rgtmlWjl/SzJ) were kindly provided by Martin Zörnig (Georg-Speyer-Haus, Frankfurt, Germany). Animals were housed at the local animal facility. High m.w. polyinosinic-polycytidylic acid [poly(I:C)] (InvivoGen, Toulouse, France) was injected i.p. at 15 μg per g body weight and mice euthanized 16 h later for isolation of splenocytes. Animal experiments were approved by the local authorities (Regierungspräsidium Darmstadt, Germany; permit numbers F146/Anz03, F146/Anz04, and FU/Anz1035) and performed in full compliance with the respective national guidelines.
Cell lines and transfectants
Cell lines were from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany). DC2.4 (48) and BWN.3G-Nkrp1g (18) reporter cells were kindly provided by Andreas Diefenbach (Mainz, Germany) and Lise Kveberg (Oslo, Norway), respectively, and maintained in RPMI 1640 supplemented either with 50 μM 2-ME, 1× MEM nonessential amino acids, and 10 μg/ml G418 (Carl Roth, Karlsruhe, Germany) (DC2.4), or with 0.5 mg/ml hygromycin B (Enzo Life Sciences, Lörrach, Germany) and 1 mg/ml G418 (BWN.3G-Nkrp1g). Chinese hamster ovary (CHO) transfectants and NK-92MI transductants were generated and cultured as described elsewhere (30). B16F10, MC38, and RAW309 cells were cultured in complete DMEM (Sigma-Aldrich, Steinheim, Germany) supplemented with 50 μM 2-ME and 1× MEM nonessential amino acids. All other cell lines were cultured in RPMI 1640 (Sigma-Aldrich).
All media contained 10% FCS (Biochrom, Berlin, Germany), 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin (all from Sigma-Aldrich).
Isolation of primary cells from mouse tissue
Single-cell suspensions of splenocytes were generated by passing the spleen through a 100 μm nylon mesh (Corning, Corning, NY) and washed with PBS supplemented with 3% FCS and 2 mM EDTA. Following lysis of erythrocytes using Pharm Lyse buffer (BD Biosciences, Heidelberg, Germany) for 3 min at room temperature, splenocytes were washed and finally passed through a 40 μm nylon mesh (Corning). For analysis of DCs or monocytes, the spleen was put into a gentleMACS C-Tube (Miltenyi Biotec, Bergisch-Gladbach, Germany) containing 2 ml HBSS buffer with Ca2+ and Mg2+49). Briefly, splenocytes were applied to a nylon wool column and nonadherent cells collected by washing the column with complete RPMI 1640 and cultured in complete RPMI 1640 supplemented with 10% FCS, 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin, 50 μM 2-ME, 1× MEM nonessential amino acids, and 1000 U/ml recombinant human IL-2 (Proleukin S; Novartis Pharma, Basel, Schweiz). Cells were passaged at days 7, 10, and 13, and nonadherent cells analyzed in all experiments. Isolation of intestinal cells was performed with modifications as described previously (50). Briefly, after removal of Peyer’s patches and longitudinal sectioning, the small intestine was washed with ice-cold PBS, cut into small pieces, and then incubated three times for 20 min at 37°C in HBSS (Sigma-Aldrich) containing 5% FCS, 5 mM EDTA, and 1 mM DTT. Each purification step was passed through a 100 μm nylon mesh, diluted with PBS followed by sedimentation of crypt cells for 10 min. A final processing through a 40 μm nylon mesh and a washing step with all pooled purification steps generated a single-cell suspension of intestinal cells. Unless stated otherwise, mice from 6 to 15 wk were used for all experiments.2, 0.5 mg/ml Collagenase D (Roche, Mannheim, Germany), and 3000 U/ml DNase I (PanReac Applichem, Darmstadt, Germany), and dissociated using the gentleMACS Octo Dissociator (Miltenyi Biotec) according to the manufacturer’s protocol. The resulting cell suspensions were then passed through a 100 μm nylon mesh and processed as described above for a single-cell suspension of splenocytes. For isolation of blood lymphocytes, blood was collected from hearts of euthanized mice, and single-cell suspensions processed as described for splenocytes. NK cells were isolated from splenocytes using the mouse NK Cell Isolation Kit II (Miltenyi Biotec) according to the manufacturer’s protocol. LAK cells were prepared as described (
Cytokine treatment, CFSE labeling, and TLR stimulation
In vitro cytokine treatment was performed with MACS-purified NK cells in complete RPMI 1640 supplemented with 10% FCS, 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin, 50 μM 2-ME and 1× MEM nonessential amino acids. Cells were seeded at a density of 1 × 106 per ml and cytokines were added at the following concentrations: 1000 U/ml (in combination with IL-15) or 10,000 U/ml (when used alone) recombinant human IL-2 (Proleukin S; Novartis Pharma), 50 ng/ml recombinant human IL-15 (Miltenyi Biotec), 2 ng/ml recombinant mouse IL-12 (PeproTech, Rocky Hill, NJ), or 50 ng/ml recombinant mouse IL-18 (MBL Pharma, Karachi, Pakistan). Stimulation was performed for the indicated time periods and cells were routinely passaged with fresh medium and cytokines at day 3 and/or the day before analysis. In long-term culture of NK cells with IL-15, the cells were passaged with fresh medium and cytokines at days 3 and 7, and then weekly. For analysis of cell proliferation, NK cells were labeled directly after purification with CFSE. Briefly, 1.5 × 106 cells were diluted in PBS/0.1% BSA (PAN Biotec, Aidenbach, Germany) at 1 × 106 6 per ml and cultured in complete RPMI 1640 supplemented with 10% FCS, 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, and 100 μg/ml streptomycin, 50 μM 2-ME, and 1× MEM nonessential amino acids. TLR agonists were used as follows: 10 μg/ml poly(I:C) (InvivoGen), 10 μg/ml LPS, and 10 μg/ml R848 (both Sigma-Aldrich) for 4–6 h.
RT-PCR and quantitative real-time PCR
Copy numbers were normalized with the ΔΔ cycle threshold method using 18S rRNA. RT-PCR for assessment of Clr-g isoform expression was performed with Phusion High-Fidelity DNA Polymerase (New England Biolabs, Frankfurt, Germany) according to the manufacturer’s protocol. cDNA was used for the PCR and was amplified in 35 cycles on a standard thermocycler. PCR products were visualized on a 5% agarose gel supplemented with ethidium bromide.
Soluble ectodomains and Clr-f/Clr-g–specific mAb 1H7
The soluble recombinant ectodomains of Nkrp1g (sNkrp1g: Pro66 through Val214) and Clr-f (sClr-f: Pro78 through Val218) were described previously (30). The Clr-f/Clr-g–specific mAb 1H7 was generated by immunizing LOU/C rats (Envigo) with the soluble ectodomain of Clr-f and resulting hybridoma were screened for specific reactivity against CHO-Clr transfectants.
Flow cytometry and cell sorting
30)], anti–Clr-f/Clr-g (clone 1H7; current study), and anti-Nkrp1g [clone 8A10 (30)]. Unless stated otherwise, all Abs are against mouse Ags. Isotype controls: mouse IgG1 (clone N1G9), rat IgG1, and rat IgG1 2a 2 (Jackson ImmunoResearch, West Grove, PA) for staining of anti-FLAG mAb M2, mouse anti–rat-A647 F(ab′)2 (Jackson ImmunoResearch) for staining of anti–Clr-f mAb 10A6, mouse anti-rat IgG1-bio (clone RG11/39.4; BD Biosciences) for staining of anti-Nkrp1g mAb 8A10 and anti–Clr-f mAb 10A8, mouse anti-rat IgM-bio (BD Biosciences) for staining of anti–Clr-f/Clr-g mAb 1H7, streptavidin-allophycocyanin and -PE (Jackson ImmunoResearch) for staining of biotinylated secondary Abs. All gatings included a singlet gating and a viability staining. Single-cell suspensions of splenocytes or NK cells were used for sorting indicated cell populations (purity >98%) using a BD FACSAria III cell sorter (BD Biosciences). For sorting of DC populations and monocytes in the resting or poly(I:C)-activated state, B cells were depleted from the single-cell suspension of splenocytes with mouse CD19 MicroBeads (Miltenyi Biotec) according to the manufacturer’s protocol.
Co-immunoprecipitation and immunoblotting
Briefly, 1 × 108 BALB/c NK cells that were expanded with IL-15 for up to 28 d were resuspended in 100 μl of PBS per stimulation and rested for 30 min at 4°C. Then cells were prewarmed for 1 min at 37°C and stimulated with 10 μl pervanadate mix (PBS with 20 mM Na3VO4, 67 mM H2O2−4% bromphenol blue) and boiling at 95°C for 10 min. Proteins were separated by SDS-PAGE, followed by wet blotting onto a nitrocellulose membrane (GE Healthcare, Little Chalfont, U.K.). Blots were blocked with TBS-T (0.1%) and 5% milk powder, and then incubated with anti–SHP-1 Ab (1:1000, clone 52/PTP1C/SHP1; BD Biosciences) overnight at 4°C. Secondary HRP-conjugated goat anti-mouse IgG Ab (1:10,000; Jackson ImmunoResearch Laboratories) was incubated for 1.5 h at room temperature and blots developed by ECL using HRP-Juice Plus (PJK, Kleinbittersdorf, Germany) and a Fusion SL machine (Vilber Lourmat, Eberhardzell, Germany).
For visualization of Nkrp1g protein in immunoblot, BALB/c or NOD SCID NK cells were stimulated for 14 d with IL-15 and then lysed using ice-cold lysis buffer. Lysis was performed for 20 min on ice and cleared by centrifugation. Then 250 μg of BALB/c, 330 μg of NOD SCID or 10 μg of NK-92MI-Nkrp1g protein lysate were deglycosylated with PNGaseF (New England Biolabs) for 1.5 h at 37°C under native conditions by only adding G7 buffer and PNGaseF. Lysates were then separated under nonreducing conditions by SDS-PAGE, followed by wet blotting onto a nitrocellulose membrane. Blocking was performed with TBS-T (0.1%) and 5% milk powder. Primary mAb 8A10-bio (30 μg/ml) was incubated overnight at 4°C, followed by HRP-conjugated streptavidin (1:25,000; Jackson ImmunoResearch Laboratories) for 1.5 h at room temperature and developed as described above.
Degranulation and cytotoxicity
Degranulation of NK-92MI-Nkrp1g transductants was quantified by flow cytometric detection of cell surface CD107a upon 1.5 h coculture with CHO transfectants at a 1:1 E:T ratio in the presence or absence of mAb 1H7. Stimulation with 25 ng/ml PMA and 1 μM ionomycin (both Sigma-Aldrich) served as positive control. NK-92MI cells were gated as CD56+ cells. Cytotoxicity was assessed by a 4 h 51Cr release assay. CHO transfectants (targets) were labeled with 51Cr for 2 h at 37°C and cocultured with NK-92MI transductants (effectors) at 37°C for 4 h at indicated E:T ratios. Coculture supernatants were mixed with OptiPhase Supermix scintillation mixture (PerkinElmer, Waltham, MA) in an IsoPlate-96 and measured with a MicroBeta2 plate counter (PerkinElmer). Percent specific lysis was calculated as follows: 100 × (experimental release − spontaneous release)/(maximum release − spontaneous release).
5 per 96 wells and incubated for 1 h at 37°C. GolgiStop (1:1500; BD Biosciences) was added to the culture. Cells were further incubated for 3 h and finally analyzed for intracellular IFN-γ expression by flow cytometry. Permeabilization and fixation was done with BD Cytofix/Cytoperm (BD Biosciences). As control, cells were only incubated with IL-15. The specific fluorescence index (SFI) of Nkrp1g+/− NK cells was calculated as follows: mean fluorescence intensity IFN-γ (IL-12, IL-18, IL-15) – mean fluorescence intensity IFN-γ (IL-15).
Reporter cell assay
BWN.3G-Nkrp1g reporter cells (1 × 104) were cocultured with poly(I:C)-activated and FACS-sorted DC populations or monocytes at a ratio of 1:2 for 18 h at 37°C in the presence or absence of mAb 1H7 in 96-well round-bottom plates. Coculture with CHO transfectants or stimulation with 25 ng/ml PMA and 1 μM ionomycin served as controls. Subsequently cells were washed and BWN.3G-Nkrp1g cells analyzed for GFP expression by flow cytometry.
NK cells were purified with mouse CD49b MicroBeads (Miltenyi Biotec) and NK cells from spleens of eight mice were sorted using FACS according to Nkrp1g expression as detected with mAb 8A10. RNA was isolated from sorted Nkrp1g+ and Nkrp1g− splenic NK cells (CD19−CD3−CD49b+NKp46+ lymphocytes), respectively, and subjected to microarray analysis (ATLAS Biolabs, Berlin, Germany) using an Affymetrix GeneChip Mouse Gene 2.0 ST array system. Three biological replicates of each cell type were analyzed. Bioinformatics analysis was performed by ATLAS Biolabs. Microarray data (accession number E-MTAB-6223) were deposited at ArrayExpress (//www.ebi.ac.uk/arrayexpress).
Statistical analyses are detailed in the figure legends and were performed with GraphPad Prism 5 (GraphPad Software, San Diego, CA).
Splenic Nkrp1g expression is confined to a small subpopulation of NK cells
We recently reported that expression of Nkrp1g receptors in the gastrointestinal tract is restricted to a subset of IEL and that this IEL-specific Nkrp1g expression is spatially matched by a selective expression of the Nkrp1g-ligand Clr-f on IEC (30). A tissue screen of BALB/c mice repeatedly exhibited in addition appreciable amounts of Nkrp1g transcripts in the spleen (Fig. 1A), which prompted us to investigate the yet undefined Nkrp1g receptor expression in the spleen. Analysis of sorted splenic lymphocyte populations by qPCR clearly attributed Nkrp1g transcripts to the NK cell population with transcripts being absent from B cells and almost undetectable in T cells (Fig. 1B). To corroborate these results, we made use of two immune-deficient mouse strains: the NOD SCID mouse, which lacks B and T cells, and the NOD SCID GAMMA mouse, which lacks B, T, and NK cells. In line with previous findings, splenocytes from BALB/c and NOD SCID mice contained substantial amounts of Nkrp1g transcripts (Fig. 1C). In contrast, spleens of NOD SCID GAMMA mice were fully devoid of Nkrp1g transcripts, consistent with the absence of NK cells in these mice. To further define the physiological expression of Nkrp1g in the spleen, which had been remained undefined mainly due to the lack of specific mAb, we assayed the expression on NK cells by flow cytometry using the recently reported Nkrp1g-specific mAb 8A10 (Fig. 1D) (30). A small NK cell subpopulation (0.5–1%) from the spleens of BALB/c mice was stained by mAb 8A10 and this staining was blocked by preincubation with soluble Nkrp1g supporting specificity (Fig. 1D). The flow cytometric staining of Nkrp1g+ NK cells by mAb 8A10 was further corroborated by using FACS to sort the 8A10-stained NK cell population followed by qPCR analysis for Nkrp1g transcripts (Fig. 1E). Sorted Nkrp1g+ γδ T cells from the small intestine (30) served as comparison. Abundance of Nkrp1g transcripts perfectly matched stainings of mAb 8A10 in splenic NK cells sorted with FACS with levels of Nkrp1g transcripts among splenic Nkrp1g+ NK cells almost identical to intestinal Nkrp1g+ γδ T cells (Fig. 1E). Additionally, NK cells that were not stained by mAb 8A10 also were nearly devoid of Nkrp1g transcripts, conclusively demonstrating specificity of 8A10 staining for Nkrp1g. Furthermore, NK cells from peripheral blood showed an Nkrp1g expression pattern similar to splenic NK cells, indicating there is a non–tissue-resident Nkrp1g+ NK cell population (Fig. 1D). Flow cytometric analyses of NK cell maturation stages as defined by CD27 and CD11b (51) did not reveal a pronounced association of Nkrp1g expression with NK cell maturation although there was a minor, but significant, increase of Nkrp1g abundance subsequent to stage I (CD27−CD11b−) (Fig. 1F).
Nkrp1g is upregulated on cytokine-stimulated NK cells
Recently, for rat Nkrp1b and Nkrp1g receptors, a modulation of surface expression upon stimulation with cytokines has been reported (52). Hence, we addressed modulation of Nkrp1g on splenic NK cells by cytokines. To this end, we isolated splenic NK cells and stimulated them in vitro with various cytokines for 7 d, and subsequently analyzed Nkrp1g expression by flow cytometry (Fig. 2A). Of note, Nkrp1g expression was strongly increased on the surface of NK cells upon treatment with either IL-2 or IL-15, with an expression 4–5 fold higher as compared with unstimulated NK cells (Fig. 2A, 2B). Nkrp1g expression was not further increased upon combined IL-2/IL-15 treatment, but was not anymore detectable upon incubation with inflammatory cytokines IL-12 and IL-18 (Fig. 2A). Flow cytometry staining of Nkrp1g by mAb 8A10 was again corroborated by FACS sorting of the respective NK cell populations and subsequent qPCR analysis of Nkrp1g transcripts, revealing that Nkrp1g transcripts exclusively reside in the NK cell subpopulation stained by 8A10 (Fig. 2C). Unexpectedly, increased Nkrp1g expression on cytokine-activated NK cells was not reflected in increased relative levels of Nkrp1g transcripts in Nkrp1g+ cytokine-activated NK cells versus Nkrp1g+ resting NK cells, indicating regulation by posttranscriptional mechanisms (Fig. 2C).
Nkrp1g homodimers are inherently expressed by a small subset of NK cells
The C-type lectin-like CLEC2 and NKRP1 family members have been described as homo- or heterodimeric disulfide-linked glycoproteins (6). Accordingly, we detected homodimers of Nkrp1g in nonreducing immunoblots of lysates of IL-15–stimulated NK cells from spleens of both BALB/c and NOD SCID mice as well as in lysates of human NK-92MI cells transduced with Nkrp1g (NK-92MI-Nkrp1g), which were included for control (Fig. 2D). Of note, there were also higher molecular forms (∼130 kDa) of Nkrp1g multimers in lysates of splenic NK cells, but not in NK-92MI-Nkrp1g controls, possibly corresponding to Nkrp1g tetramers or Nkrp1g heteromers. Next, we addressed inherent stability of Nkrp1g expression by NK cells. NK cells were isolated from spleens of BALB/c mice, stimulated with IL-15 for 7 d in vitro, and then sorted via FACS according to Nkrp1g expression using mAb 8A10. Subsequently, the Nkrp1g+ and Nkrp1g− populations were separately stimulated for another 7 d with IL-15 and then analyzed for Nkrp1g expression by flow cytometry (Fig. 2E). Remarkably, the Nkrp1g− NK cell fraction remained fully negative for Nkrp1g, whereas the Nkrp1g+ mostly maintained the initial Nkrp1g expression (96.9% after sorting; 87.2% after 7 d IL-15). A differential proliferation of both subsets was addressed by a CFSE proliferation assay (Fig. 2F), which did not reveal any significant difference in proliferation of Nkrp1g+ versus Nkrp1g− NK cells after 3 or 7 d, respectively. In conclusion, Nkrp1g expression appears to be stably associated with a small subset of splenic NK cells, pointing to an inherent Nkrp1g expression phenotype of these NK cells.
Strain-specific expression of Nkrp1g on BALB/c NK cells
Intestinal Nkrp1g expression on IELs has been reported to be strain specific, because there are ∼10% Nkrp1g+ γδ IEL in BALB/c mice, whereas only ∼1% of γδ IEL of C57BL/6 mice express Nkrp1g at the cell surface (30). In this study, we observed a similar strain-associated bias for Nkrp1g expression by splenic NK cells with frequencies of Nkrp1g+ NK cells being about four times higher among BALB/c NK cells as compared with C57BL/6 NK cells (∼0.8% versus 0.2%) (Fig. 3A). However, Nkrp1g expression in C57BL/6 mice does not appear to be restricted transcriptionally, as Nkrp1g transcript expression is five times higher in splenocytes of C57BL/6 mice compared with splenocytes of BALB/c mice (Fig. 3B), although frequencies of NK cells in C57BL/6 spleens were lower than in BALB/c (BALB/c: ∼3.2% NK cells, C57BL/6: ∼1.6% NK cells). Amino acid sequences of Nkrp1g differ in both mouse strains by a conservative substitution of the carboxyterminal amino acid (Ile in BALB/c versus Val in C57BL/6) (Fig. 3C), which is unlikely to explain the marked differences in Nkrp1g+ NK subsets by altered protein expression or a differential recognition by mAb 8A10. Rather, an additional regulatory mechanism may differentially control Nkrp1g surface expression in BALB/c versus C57BL/6 mice. Low frequencies of Nkrp1g+ NK cells (<1%) were also found among splenic NK cells from inbred mouse strains C3H and 129 as well as from outbred mouse strain CD-1 (data not shown), suggesting that rareness of Nkrp1g+ splenic NK cells is a general feature of Mus musculus.
Characterization of Nkrp1g+ NK cells
To more precisely define the Nkrp1g+ NK cell subset, we performed flow cytometry stainings for several NK cell markers and receptors and found that Nkrp1g+ NK cells are more skewed toward expressing the inhibitory Ly49C receptor on their surface (Fig. 4A, 4B). Further characterization of Nkrp1g+ NK cells was carried out by microarray analysis of FACS sorted Nkrp1g+ versus Nkrp1g− splenic NK cells. We found 214 differentially regulated genes that were at least 2-fold up- or downregulated (38 upregulated versus 176 downregulated). A selection of differentially expressed immune-related genes is displayed in Fig. 4C. Among the downregulated genes were transcription factors T-bet and Rxra, the CLEC2 family member and Nkrp1b ligand Clr-b (Clec2d) as well as S1pr1. However, no significant differences were detected for transcription factors Eomes, GATA3, and RORγT, indicating that Nkrp1g+ NK cells are bona fide NK cells. Genes upregulated in Nkrp1g+ NK cells were Ly49e and the decay accelerating factor CD55. The latter was also found upregulated in Nkrp1g+ γδ IEL (30), linking expression of Nkrp1g and CD55 in a yet unknown fashion. GO enrichment analysis of downregulated genes revealed several enriched gene ontology terms, most notably innate immune response in mucosa and lymphocyte migration, which may possibly indicate a less migratory property (Supplemental Table I). Nkrp1f shares two ligands with Nkrp1g (Clr-d and Clr-g), and therefore has been considered as the complementary receptor for Nkrp1g. Hence, coexpression of Nkrp1f on Nkrp1g+ NK cells was of particular interest. Analyzing Nkrp1f transcripts in subsets of lymphocytes and DCs clearly confines Nkrp1f expression to NK cells (Fig. 4D). However, Nkrp1f surface expression on Nkrp1g+ NK cells did not significantly vary from the previously described broad expression of Nkrp1f on the vast majority of splenic NK cells (Fig. 4E) (28). Finally, to address functional differences between Nkrp1g+ and Nkrp1g− NK cells, we assessed degranulation and IFN-γ secretion. We consistently observed that IL-15–stimulated Nkrp1g+ NK cells are producing significantly more IFN-γ after exposure to IL-12 plus IL-18 than Nkrp1g− NK cells (Fig. 4F), whereas no differences were observed with regard to degranulation in response to several target cells (data not shown).
Differential Nkrp1g ligand expression in the splenic compartment
Nkrp1g ligates three CLEC2 members, namely Clr-d, Clr-f, and Clr-g (18, 29). A qPCR-based tissue screen of BALB/c mice for Clr-f transcripts revealed abundant amounts of Clr-f transcripts in the small intestine, and, at lower levels, in the colon and kidney, as previously described (30). However, Clr-f transcripts were barely detectable or undetectable in all other tissues analyzed including spleen and lymph nodes (Fig. 5A). Because a previous study had described Clr-f transcripts in LAK cells (5), we stained various splenic leukocyte populations, cytokine-stimulated NK cells, and LAK cells with the Clr-f-specific mAb 10A6 and 10A8 (30), but could not detect any Clr-f surface expression (Supplemental Fig. 1). Further, we assessed levels of Clr-f transcripts in LAK cells and IL-15–stimulated NK cells and found that they were more than 10,000-fold lower than in the small intestine, explaining the lack of cell surface expression (Fig. 5B). Subsequently, we performed qPCR-based tissue screens of BALB/c mice for transcripts of the other Nkrp1g-ligands Clr-d and Clr-g, respectively. Although Clr-d transcripts are most abundant in the eye, and at lower levels detectable in lung, small intestine, colon, thymus, and kidney, they were hardly detectable in the spleen (Fig. 5C). In stark contrast, Clr-g transcripts are most abundant in lymphoid organs such as the spleen and the thymus, strongly arguing for a preferential expression by hematopoietic cells (Fig. 5D).
From these qPCR-based tissue screens, Clr-g was singled out as putative ligand of splenic Nkrp1g+ NK cells, because both Clr-d and Clr-f transcripts were virtually absent from the spleen. The splenic compartment was further fractionated by FACS sorting into various leukocyte populations, which were analyzed by qPCR for Clr-g transcripts. It turned out that Clr-g transcripts are broadly present in both splenic lymphocytes (T cells, B cells, NK cells) and myeloid cells (monocytes, DCs) at similar levels (Fig. 5E). Next, we assessed the abundance of Clr-g transcripts in various mouse cell lines (Fig. 5F). All tested T cell lines (YAC-1, RMA, RMA-S, EL4) and B cell lines (Ba/F3, SP2/0) contained Clr-g transcripts at levels roughly equivalent to splenic lymphocytes. Of note, the NK sensitive T cell line YAC-1 contained the lowest levels of Clr-g transcripts, whereas both B cell lines exhibited the highest levels in line with highest transcript levels found in splenic B cells. However, nonlymphocytic hematopoietic cell lines (P815, RAW309, DC2.4) as well as cell lines of nonhematopoietic origin (B16F10, MC38) were fully devoid of Clr-g transcripts (Fig. 5F).
Converse Clr-g surface expression on cell lines and splenic leukocytes
Detection of abundant Clr-g transcripts in leukocytes as well as in T and B cell lines then prompted us to characterize Clr-g surface expression by flow cytometry. By immunizing rats with recombinant soluble Clr-f we had generated mAb 1H7 that detects both Clr-f and Clr-g, but does not cross-react with Clr-a, Clr-b, or Clr-c, respectively (Supplemental Fig. 2). Hence, we performed flow cytometry with mAb 1H7 always in parallel with the Clr-f specific mAb 10A8 to address a potential confounding effect of Clr-f expression. However, mAb 10A8 did not bind to any of the cell lines tested nor to any splenic subpopulation well in line with the complete lack of Clr-f transcripts (Supplemental Fig. 1). In contrast, mAb 1H7 strongly stained all T cell lines (YAC-1, RMA, RMA-S, EL4) and B cell lines (Ba/F3, SP2/0) shown to abundantly contain Clr-g transcripts, whereas all other cell lines, which are devoid of Clr-g transcripts (P815, RAW309, DC2.4, B16F10, MC38), also did not bind mAb 1H7, establishing a perfect match between 1H7 binding and abundance of Clr-g transcripts (Fig. 6A).
We next analyzed splenic B cells, T cells, NK cells, pDCs, CD8α+/− cDCs, and Ly6Chi/lo monocytes for Clr-g surface expression by flow cytometry. Unexpectedly, mAb 1H7 did not detect any surface Clr-g on B cells, T cells, NK cells, or DC populations, despite Clr-g transcript levels comparable to T and B cell lines. Only monocytes weakly bound mAb 1H7 (Fig. 6B). Previously, we described that Clr-f is upregulated on IECs after challenge with poly(I:C) (30). Therefore, we wondered whether Clr-g surface expression may become detectable upon in vivo challenge with poly(I:C). Although monocytes and lymphocytes from spleens of poly(I:C)-treated mice were not stained, there was a strong and specific induction of Clr-g surface expression on CD8α+ cDCs (Fig. 6C). To validate these results, we analyzed FACS sorted subsets of splenic DCs and monocytes for Clr-g transcripts. In line with the results above, Clr-g transcripts were detected in all subsets of DCs and monocytes from naive mice at roughly similar levels (Fig. 6D). However, analysis of these subsets from poly(I:C)-treated mice, revealed a pronounced selective induction of Clr-g transcripts in CD8α+ cDCs (∼35-fold) well in line with the observed upregulation of Clr-g on the cell surface.
We considered the possibility that alternative splicing and expression of different Clr-g isoforms in splenic lymphocytes versus T and B cell lines may account for the observed discrepancy between Clr-g transcript levels and Clr-g surface expression. There are four isoforms of differentially spliced Clr-g transcripts (Supplemental Fig. 3A). Using a PCR-based assay distinguishing all four isoforms, we found that Clr-g transcript isoform 2 is uniformly and predominantly expressed in all tested DCs, monocytes, and lymphocytes, as well as in Clr-g–expressing cell lines (Supplemental Fig. 3B, 3C), suggesting that alternative splicing of Clr-g transcripts is not accountable for the deficient Clr-g surface expression on lymphocytes.
Clr-g surface expression on CD8α+ cDCs is type I IFN receptor dependent
CD8α+ cDCs represent the major DC population expressing TLR3 and are hence directly responsive to stimulation with poly(I:C), a mimic of viral dsRNA. As we had observed a poly(I:C)-induced surface expression of Clr-g on CD8α+ cDCs in vivo after 16 h, we wondered if this could be recapitulated in vitro. To this end, we stimulated splenocytes in vitro with different TLR ligands [poly(I:C), LPS, R848] for 4 h (Fig. 7A). Interestingly, only R848, a ligand of TLR7/8, was able to induce Clr-g surface expression on the pDC population in vitro after this short timeframe. To further analyze the kinetic of poly(I:C)-induced Clr-g expression on CD8α+ cDCs, we treated mice with poly(I:C) and harvested spleens at different time points to stain for Clr-g expression (Fig. 7B). Clr-g surface expression on CD8α+ cDCs could be only detected after 16 h but not at earlier time points, suggesting that Clr-g might not be directly induced upon TLR3 stimulation. A previous study has shown that signaling by the type I IFN receptor (IFNAR) plays an indispensable role in inducing immunogenicity of DCs after poly(I:C) treatment by analyzing gene expression of wild-type DCs or IFNAR−/− DCs upon poly(I:C) injection (53). Analyzing this available dataset for Clr-g expression (gene Clec2i) revealed that induced Clr-g expression by DCs upon poly(I:C) treatment was strongly dependent on IFNAR (Fig. 7C), supporting the assertion of Clr-g being secondarily induced upon TLR3 stimulation.
Nkrp1g and Clr-g interaction is inhibitory
To address functional consequences of an interaction between Nkrp1g and Clr-g, we performed a degranulation assay using NK-92MI-Nkrp1g transductants together with CHO-Clr-g transfectants or control-transfected CHO cells (Fig. 8A). Degranulation of NK-92MI-Nkrp1g cells in cocultures with CHO–Clr-g cells was strongly reduced as compared with cocultures with CHO-mock cells, but could partially be restored when Clr-g–specific mAb 1H7 was added to the culture (Fig. 8A). In chromium release assays, CHO–Clr-g and CHO-mock cells were similarly lysed by control-transduced NK-92MI cells, whereas cytolysis of CHO–Clr-g by NK-92MI-Nkrp1g was selectively abrogated as compared with CHO-mock cells (Fig. 8B). These experiments showed that Nkrp1g engagement by Clr-g effectively inhibits effector functions of NK-92 cells. This is likely mediated through ITIM signaling as Nkrp1g bears an ITIM motif in its cytoplasmic tail (29). Inhibitory receptors bearing an ITIM in their cytoplasmic tail become phosphorylated by Src kinases upon activation and allow recruitment of phosphotyrosine phosphatases such as SHP-1 as shown for inhibitory Nkrp1b (54, 55). We therefore attempted to coimmunoprecipitate SHP-1 with Nkrp1g from IL-15–expanded BALB/c NK cells (Fig. 8C). Immunoblotting detected SHP-1 in Nkrp1g-immunoprecipitates, independently of a pretreatment with pervanadate, pointing to a constitutive association of SHP-1 with Nkrp1g.
Finally, we addressed whether Clr-g induced on CD8α+ cDCs upon poly(I:C) challenge can be recognized via Nkrp1g. To this aim, pDCs, CD8α+ cDCs, CD8α− cDCs, and monocytes were FACS sorted from splenocytes of poly(I:C)-treated mice and cultured together with BWN-Nkrp1g transfectants in the absence or presence of mAb 1H7. BWN-Nkrp1g reporter cells specifically respond to CHO–Clr-g cells, but not to CHO-mock controls, and the response to CHO–Clr-g can be specifically blocked by addition of mAb 1H7 to the cocultures (Fig. 8D). Although BWN-Nkrp1g reporter cells did not specifically react in cocultures with pDCs, CD8α− cDCs, and monocytes, they specifically responded to CD8α+ cDCs in a Clr-g–specific manner as the response could be blocked by the addition of mAb 1H7 (Fig. 8D).
Understanding immune cell activation and regulation is indispensable when it comes to the treatment of infectious diseases or cancer. NK cells in this study play an important role by eradicating infected or transformed cells, which is mediated through a plethora of germline-encoded activating and inhibitory receptors (3, 56, 57). However, NK cells not only eliminate dangerous cells through cytotoxic activity, they also shape the adaptive immune response by killing activated lymphocytes and engaging in a mutual cross-talk with DCs (58, 59). Both soluble factors and cell-cell contacts contribute to this NK-DC cross-talk. For example, cytokines such as IL-12, -15, and -18, and type I IFNs from activated DCs can stimulate IFN-γ secretion and cytotoxicity of NK cells (47, 60, 61). In turn, TNF or IFN-γ secreted by NK cells promote the maturation of DCs and thereby influence the priming of T cell responses. In addition, interactions of NK cell receptors such as NKp30, CD94/Nkg2a, Nkg2d, and CD27 with their respective ligands on DCs reportedly shape the DC repertoire (62–67). It has also been shown that Clr-b and Nkrp1b play a role in NK-DC interactions (8, 68).
In this study we report an exclusive expression of the C-type lectin-like receptor Nkrp1g on a minor subset of splenic NK cells, which is matched by selective expression of the Nkrp1g ligand Clr-g on activated splenic DCs. Previous studies indicated the presence of Nkrp1g transcripts in the intestine, spleen, and thymus of various strains of mice (30, 31), but cellular Nkrp1g expression as well as splenocytes expressing Nkrp1g receptors remained undefined. Recently, we showed that in the small intestine, Nkrp1g is expressed on subsets of IEL, particularly on γδ IEL (30). In contrast, we find Nkrp1g in the spleen exclusively expressed by a minor subset of NK cells, which is also present in the periphery (Fig. 1D). Nkrp1g expression on resting NK cells is rather weak, but is markedly upregulated upon exposure to cytokines IL-2 and/or IL-15. This is well in line with a previous study on rat NK cells showing upregulation of Nkrp1g in response to IL-2 treatment (52). This study also reported an upregulation of Nkrp1g on rat NK cells upon IL-18 treatment. However, we observed a loss of Nkrp1g+ mouse NK cells following exposure to IL-18, pointing to a differential regulation of Nkrp1g in rat and mouse. In further contrast to mice, Nkrp1g is expressed more broadly in rats, with an expression on 10% NK cells and 13% NK-like T cell subsets (52). Nkrp1g expression of mouse NK cells appears to be quite stable and limited to a small subset, pointing to an epigenetically fixed expression pattern. However, Nkrp1g surface expression is regulated in a strain-specific manner, as Nkrp1g+ IELs and Nkrp1g+ splenic NK cells are almost absent in C57BL/6 mice as compared with BALB/c mice, although Nkrp1g transcripts are present at similar levels. This points to a differential posttranscriptional regulation of Nkrp1g expression in these mouse strains. Similarly, Nkrp1g transcript levels in cytokine-stimulated Nkrp1g+ NK cells were lower than in resting Nkrp1g+ NK cells despite high Nkrp1g surface expression on activated NK cells, suggesting an additional level of regulation.
Characterization of Nkrp1g+ NK cells revealed an increased coexpression of the inhibitory Ly49C, indicating a population of NK cells that is licensed and controlled by MHC class I–specific receptors. Of note, we also found that Nkrp1g expression in both IEL and splenic NK cells is linked to an enhanced expression of CD55, which is protective against lysis by the complement system. Nkrp1f, which shares ligands Clr-d and Clr-g with Nkrp1g, was not differentially expressed between Nkrp1g+ and Nkrp1g− cells in line with the findings in rats, where Nkrp1f is also coexpressed on NK cells (52).
By qPCR-based tissue screens we singled out Clr-g as a putative splenic Nkrp1g ligand, as Clr-d and Clr-f transcripts were almost absent from the spleen. Our data further support the idea that most Clr molecules are differentially expressed in a tissue-specific or tissue-biased manner, and thus may be involved in a tissue-specific immunosurveillance (30, 31). The only exception thus far is Clr-b, which is expressed rather ubiquitously like MHC class I (30, 31), and ligates the inhibitory receptor Nkrp1b presumably acting as an MHC class I–independent missing-self detection system (19, 23, 25, 26). In contrast, expression of the three Clr molecules ligated by the inhibitory Nkrp1g receptor is strongly biased toward specific cell types: Clr-f expression is restricted to the gastrointestinal epithelium and the kidney, Clr-d is preferentially expressed in the eye, whereas Clr-g expression appears to be associated with cells of hematopoietic origin. Such a tissue-biased expression pattern of Clr molecules might explain the multiple ligand specificities of Nkrp1g and Nkrp1f, respectively, by specifically regulating the activity of NK and/or T cells depending on the tissue context.
Analyzing Clr-g transcripts in mouse splenocytes revealed a broad presence in all tested subpopulations, whereas mouse cell lines showed a more restricted Clr-g expression primarily by mouse T and B cell lines. Using a novel anti–Clr-g mAb, we could show that the presence of Clr-g transcripts in these cell lines perfectly matches Clr-g surface expression. In contrast, no Clr-g molecules were detectable on splenic lymphocytes or DCs, despite transcript levels comparable to T and B cell lines, with only residual Clr-g expression on splenic monocytes. This points to an additional posttranscriptional mechanism stringently controlling Clr-g surface expression in lymphocytes that is absent in tumor cell lines. This is also reminiscent of the restricted surface expression of the human CLEC2 family member and NKp80 ligand AICL, which is only residually expressed on human monocytes despite abundant transcripts in all leukocyte populations (12). Indirect evidence for a low or absent Clr-g expression on primary splenocytes was already provided by an early study, where Nkrp1f reporter cells did not respond to mouse splenocytes (8). In this study, Nkrp1f reporter cells selectively responded against cultured bone marrow–derived DC and bone marrow–derived macrophages, and the response was abrogated upon LPS treatment, suggesting LPS-mediated downregulation of the respective Nkrp1f ligand. However, it remained unclear which Clr molecule is recognized by Nkrp1f in this context. In contrast, we observed a marked upregulation of Clr-g on the surface of splenic DCs upon challenge with TLR ligands. We found that Clr-g is upregulated on CD8α+ cDCs upon in vivo challenge with poly(I:C), which is in line with an upregulation of Clr-g transcripts in this cell population. Such in vivo–induced Clr-g expression on DCs could be sensed by Nkrp1g as shown by Nkrp1g-expressing reporter cells. In vitro, we observed that Clr-g surface expression can be rapidly induced on pDCs upon treatment with R848, which is in line with the expression of TLR7/9 by pDCs. However, we were unable to recapitulate the poly(I:C)-induced Clr-g induction on CD8α+ cDCs in vitro due to the rapid apoptosis of splenic DCs in vitro (69), which precluded analyses of Clr-g induction after extended in vitro culture. The delayed Clr-g induction in vivo upon poly(I:C) challenge, together with data from a previous study addressing the importance of IFNAR signaling for poly(I:C)-induced DC maturation (53), strongly suggest that Clr-g upregulation is dependent on IFNAR signaling following poly(I:C) challenge. IFNAR dependency was also recently shown for Clr-b, which was upregulated after MCMV infection on bystander cells in a type I IFN–dependent manner (70). To provide a rationale for the differential kinetics of delayed Clr-g upregulation after poly(I:C) challenge, which is dependent on IFNAR signaling, versus the rapid R848-mediated Clr-g upregulation, we performed an in silico analysis of the Clr-g promotor for transcription factor binding sites [MatInspector, Genomatix (71)]. Although no binding site for IRF3, which signals downstream of TLR3, was predicted, a putative binding site of IRF7 was identified, which mediates signals downstream of TLR7/9.
Taken together, Clr-g is preferentially, if not selectively, expressed by hematopoietic cells and Clr-g surface expression appears tightly regulated in splenocytes by a yet unknown posttranscriptional mechanism. Hence, Clr-g surface expression is mostly absent from resting splenocytes, but can be markedly induced on DCs upon challenge with TLR ligands. Such induced Clr-g expression on DCs could be sensed by a small subset of splenic NK cells inherently expressing the inhibitory Nkrp1g receptor. It has been previously shown that NK cells and DCs colocalize in the red pulp of the spleen in the resting state (72). Upon inflammation induced by poly(I:C), LPS or MCMV infection, NK cells can leave the red pulp or travel to T cell areas of the white pulp (73) where mostly CD8α+ cDCs are located (74). Hence, an interaction between Nkrp1g on activated splenic NK cells and Clr-g on stimulated DCs may occur in such a scenario. Because Clr-g is also a ligand of the more broadly expressed activating Nkrp1f receptor, it remains to be determined how such an inhibitory Nkrp1g–Clr-g interaction interferes in functional terms with a parallel engagement of Nkrp1f. Although our preliminary data indicate that in vivo activated CD8α+ cDCs engage Nkrp1f ectodomains and ectopically expressed Nkrp1f triggers NK cell degranulation upon Clr-g engagement (data not shown), the functional consequences of Nkrp1g/Nkrp1f interactions with Clr molecules need to be assessed in more physiological settings that are hindered by the complexity of Clr ligands and scarcity of Nkrp1g expressing cells. Whereas engagement of Nkrp1g by Clr-g dampens NK cell responses through inhibitory signaling via SHP-1, a reverse signaling by Clr-g is unlikely as its cytoplasmic portion is devoid of signaling motifs and Clr-g lacks charged residues in the transmembrane region required for association with signaling adapters. Altogether, our study adds to our knowledge a potential NK-DC receptor-ligand interaction that may be involved in regulating the immune response after infection.
The authors have no financial conflicts of interest.
We thank Christian Lehmann for critical advice in DC phenotyping, Tobias Zöller for help in cytokine detection, and Praveen Mathoor for cell sorting.
This work was supported in part by grants to D.D. from the German Research Foundation (CRC1181-TPA7) and the Interdisziplinäres Zentrum für Klinische Forschung (IZKF-A65).
The microarray data presented in this article have been submitted to ArrayExpress (//www.ebi.ac.uk/arrayexpress) under accession number E-MTAB-6223.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- conventional DC
- Chinese hamster ovary
- C-type lectin-like 2
- dendritic cell
- intestinal epithelial cell
- intestinal intraepithelial lymphocyte
- type I IFN receptor
- lymphokine-activated killer
- murine CMV
- NK gene complex
- plasmacytoid DC
- polyinosinic-polycytidylic acid
- quantitative PCR
- specific fluorescence index.
- Received August 15, 2017.
- Accepted November 13, 2017.
- Copyright © 2018 by The American Association of Immunologists, Inc.