Abstract
Natural interferon (IFN)-producing cells are the primary cell type responsible for production of type I IFN in response to viruses. Herein we report the identification of the first molecular marker of mouse natural interferon-producing cells (IPCs), a novel member of the sialic acid-binding immunoglobulin (Ig)-like lectin (Siglec) family termed Siglec-H. Siglec-H is expressed exclusively on IPCs and is unique among Siglec proteins in that it associates with the adaptor protein DAP12. Moreover, we show that DAP12 modulates the type I IFN response of IPCs to a Toll-like receptor 9 (TLR9) agonist. This observation explains our previous finding that stimulation of IPCs with 440c, a Siglec-H-specific antibody, reduces IPC secretion of type I IFN. Moreover, it supports a model in which engagement of DNAX-activation protein 12 (DAP12)-associated receptors with antibodies or low avidity endogenous ligands interferes with TLR-mediated cellular activation. (Blood. 2006;107:2474-2476)
Introduction
Natural interferon-producing cells (IPCs), also called plasmacytoid dendritic cells, are a rare subset of cells that are thought to play a key role in antiviral responses.1 Through Toll-like receptor 9 (TLR9) and TLR7, IPCs detect DNA and RNA viruses2 and respond by secreting high levels of type I interferons (ie, IFN-α and IFN-β),3-5 interleukin 12 (IL-12),4-6 and proinflammatory chemokines.1 In the mouse, IPCs have a complex phenotype (CD11c+, Ly-6C+, B220+, CD11b-, CD8a+/-),5 making their characterization challenging. To facilitate mouse IPC identification, we recently established a monoclonal antibody (mAb), 440c, which recognizes an IPC-restricted surface marker.7 Binding of mAb 440c results in decreased IFN-α production in vitro in response to TLR9 ligands, such as CpG oligonucleotides. In vivo, treatment of mice with mAb 440c prior to CpG treatment reduces systemic IFN-α production but does not deplete IPCs. Thus, mAb 440c inhibits IFN-α secretion by IPCs in vitro and in vivo. The molecule recognized by mAb 440c and the mechanism by which it modulates IPC function is unknown. Here we demonstrate that mAb 440c recognizes a previously uncharacterized member of the sialic acid-binding immunoglobulin (Ig)-like lectins (Siglec) family, Siglec-H; that Siglec-H associates with the adaptor DAP12 for signaling; and that DAP12 signaling reduces the type I IFN response of IPCs to a TLR9 agonist.
Study design
Flow cytometry
IFN-α assays
Splenic CD11c+B220+CD11b-Ly6C+ IPCs from wild-type (WT) C57BL/6 and DAP12-/- mice were sorted by flow cytometry and stimulated with CpG oligonucleotide 2216 for 16 hours. Mice received intravenous injections of 5 μg CpG 2216 in liposomal transfection reagent (DOTAP; Roche, Indianapolis, IN). IFN-α was measured in cell culture supernatants or mouse serum by enzyme-linked immunosorbent assay (ELISA; PBL Biomedical Laboratories, New Brunswick, NJ). Statistical analysis was performed with Graphpad Prism Software.
Expression cloning
IPCs were cultured from bone marrow cells of C57BL/6 mice10,11 and isolated by magnetic cell sorting (MACS, Miltenyi Biotec, Auburn, CA). IPC RNA was extracted by the Triazol method (Invitrogen, Carlsbad, CA) and used to prepare a customized cDNA library in Express-1 (Open Biosystems, Huntsville, AL). The IPC cDNA library was transfected in 293T cells stably expressing DAP12. Transfected cells expressing 440c antigen (Ag) were sorted by flow cytometry.
Results and discussion
DAP12-deficient IPCs lack expression of 440c
DAP12 is a transmembrane adaptor molecule that is required for cell surface expression and signaling of many immune receptors expressed in natural killer (NK) cells, macrophages, and dendritic cells (DCs).12 To determine whether DAP12 deficiency altered IPC development and/or function, we stained cell suspensions of spleens derived from DAP12-deficient mice with mAbs 440c and mPDCA1, which recognize distinct IPC-specific markers (Blasius et al,7 Krug et al,13 and data not shown). Although DAP12-/- spleens contained IPCs expressing the mPDCA1 Ag, these IPCs lacked the 440c Ag (Figure 1A). In contrast, the IPCs from mice deficient for other signaling adaptors, such as DAP10 and the gamma chain of Fc receptors (FcRγ), expressed both 440c and mPDCA1 antigens (Figure 1A), indicating that mAb 440c recognizes a DAP12-associated receptor.
Antibody 440c recognizes Siglec-H. (A) Total splenocytes from WT (C57BL/6), DAP12-/-, DAP10-/-, or FcRγ-/- mice (all on C57BL/6 background) were stained with phycoerythrin (PE)-labeled mAb anti-B220 and biotinylated mAbs 440c7 or mPDCA1,13 followed by streptavidin-APC and analyzed by 2-color flow cytometry. Percentages of 440+ and mPDCA1+ IPCs are indicated. IPCs appear to be less abundant in WT mice than in DAP12-/- mice. It is possible that the lack of 440c Ag expression in the DAP12-/- mice influences IPC homing to the spleen, IPC development, or IPC survival. (B) 293T cells expressing murine DAP12 (293T-DAP12) and 293T cells were transiently transfected with a Siglec-H-encoding cDNA, stained with mAb 440c, and analyzed by flow cytometry. FSC, forward scatter. (C) Alignment of V-type Ig domains of Siglec-H and mouse CD33. Identical residues are darkly shaded, similar residues are lightly shaded. Residues likely involved in sialic acid binding are boxed.16 (D) Alignment of transmembrane regions of mouse Siglec-H, mouse CD33, and chimpanzee Siglec-13. The conserved lysine residue that may allow for DAP12 association is indicated with an asterisk.
Antibody 440c recognizes Siglec-H. (A) Total splenocytes from WT (C57BL/6), DAP12-/-, DAP10-/-, or FcRγ-/- mice (all on C57BL/6 background) were stained with phycoerythrin (PE)-labeled mAb anti-B220 and biotinylated mAbs 440c7 or mPDCA1,13 followed by streptavidin-APC and analyzed by 2-color flow cytometry. Percentages of 440+ and mPDCA1+ IPCs are indicated. IPCs appear to be less abundant in WT mice than in DAP12-/- mice. It is possible that the lack of 440c Ag expression in the DAP12-/- mice influences IPC homing to the spleen, IPC development, or IPC survival. (B) 293T cells expressing murine DAP12 (293T-DAP12) and 293T cells were transiently transfected with a Siglec-H-encoding cDNA, stained with mAb 440c, and analyzed by flow cytometry. FSC, forward scatter. (C) Alignment of V-type Ig domains of Siglec-H and mouse CD33. Identical residues are darkly shaded, similar residues are lightly shaded. Residues likely involved in sialic acid binding are boxed.16 (D) Alignment of transmembrane regions of mouse Siglec-H, mouse CD33, and chimpanzee Siglec-13. The conserved lysine residue that may allow for DAP12 association is indicated with an asterisk.
Antibody 440c recognizes Siglec-H
To identify the DAP12-associated molecule recognized by mAb 440c, 293T cells stably expressing mouse DAP12 (293T-DAP12) were transiently transfected with a mouse IPC cDNA library. mAb 440c-positive cells were sorted by flow cytometry. cDNA clones extracted from 440c+ 293T-DAP12 cells were subjected to several rounds of transfection and sorting, until an individual cDNA clone was obtained that conferred mAb 440c binding in 293T-DAP12 cells but not 293T cells (Figure 1B). Sequence analysis of the isolated clone revealed that mAb 440c recognizes a member of the Siglec family, which has been designated Siglec-H.14
Siglec-H exhibits structural features unique to DAP12-associated receptors
Siglecs are transmembrane glycoproteins of the Ig superfamily that bind sialic acids decorating a variety of glycoproteins and glycolipids.14,15 Siglecs include Sialoadhesin (Sn), CD22, myelin-associated glycoprotein (MAG), CD33, and various CD33-related Siglecs. All Siglecs bind sialic acids through their N-terminal V-type Ig domain, which is followed by varying numbers of C2-type Ig domains. Siglec-H structurally resembles CD33, as it includes a V-type domain followed by a single C2-type Ig domain. The IgV-type domain of Siglec-H contains the structural features that have been shown to be required for sialic acid recognition, particularly a conserved arginine residue in the F strand that forms a salt bridge with the carboxylate of sialic acid (Figure 1C).14-16 This suggests that Siglec-H binds sialylated ligands.
In the CD33-related Siglecs characterized thus far, the cytoplasmic domains include immune receptor tyrosine-based inhibitory motifs (ITIMs), which recruit the inhibitory protein tyrosine phosphatases SHP-1 and SHP-2.14,15 In remarkable contrast with the other CD33-related Siglecs, Siglec-H lacks cytoplasmic ITIMs and contains a charged lysine residue in the transmembrane region (Figure 1D). A transmembrane lysine residue has been previously shown to allow pairing of other cell surface receptors with a charged aspartic acid residue in the transmembrane region of DAP12. Thus, this structural feature of Siglec-H is consistent with its association with DAP12. Of note, Siglec-H is the first Siglec found to be associated with DAP12. Another murine Siglec, CD33, and the chimpanzee Siglec 13 contain a transmembrane lysine and are likely to also associate with DAP12 (Figure 1D). However, none of the 11 human Siglecs appear to associate with DAP12, indicating that Siglec-H may have rapidly evolved under selective pressure, perhaps for resistance to a defined infectious agent.
DAP12-deficient IPCs secrete more IFN-α than WT IPCs
We previously showed that mAb 440c inhibits IFN-α secretion of IPCs in response to CpG oligonucleotides.7 This observation is apparently in contrast with our observation that Siglec-H associates with DAP12, as DAP12 contains a cytoplasmic ITAM, which triggers activating signals by recruiting protein tyrosine kinases.12 Since Siglec-H is the only DAP12-associated receptor found on mouse IPCs, to address this discrepancy we compared IFN-α secretion of DAP12-deficient and WT IPCs in response to CpG in vivo and in vitro. DAP12-deficient IPCs secreted more IFN-α than WT IPCs after in vitro stimulation with low concentrations of CpG (Figure 2A). At high CpG concentration, DAP12-deficient and WT IPCs secreted similar amounts of IFN-α (Figure 2A). Moreover, following intravenous injection of CpG, serum levels of IFN-α were significantly higher in DAP12-deficient mice than in WT mice (Figure 2B). These results indicate that DAP12 reduces the responsiveness of IPCs to CpG. This conclusion is corroborated by our recent observation that engagement of a human DAP12-associated receptor on IPCs, NKp44, also inhibits IFN-α secretion.17 Moreover, DAP12-deficient bone marrow-derived macrophages were shown to be more responsive to low concentrations of TLR ligands than WT macrophages.18
In conflict with these observations, biochemical studies demonstrate that phosphorylation of the DAP12 ITAM leads to recruitment and activation of signaling mediators canonically associated with cellular activation.12 To reconcile these data, we suggest a model in which tonic low-level activation of DAP12 by low-avidity endogenous ligands18 or certain monoclonal antibodies results in a partial activation of DAP12, sufficient to recruit activating signaling mediators, but insufficient to mediate downstream signaling. These signaling mediators are thus functionally sequestered with DAP12 and are unable to participate in downstream TLR-mediated cellular activation, leading to a net inhibition in TLR signaling by DAP12 activation.
DAP12-deficient IPC secrete more IFN-α than WT IPC in vitro and in vivo. (A) Splenic CD11c+B220+CD11b-Ly6C+ IPCs from WT C57BL/6 and DAP12-/- mice were sorted by flow cytometry. Cells were stimulated with CpG oligonucleotide 2216 for 16 hours. Data are compiled from 6 independent experiments. (B) Mice received intravenous injections of 5 μg CpG 2216 in DOTAP and serum was collected after 6 hours. Representative data of 1 of 3 independent experiments are shown, each with 3 mice per group. IFN-α was measured in cell culture supernatants (A) or mouse serum by ELISA (B). Error bars show one standard deviation. *P < .05 versus B6 by Student t test.
DAP12-deficient IPC secrete more IFN-α than WT IPC in vitro and in vivo. (A) Splenic CD11c+B220+CD11b-Ly6C+ IPCs from WT C57BL/6 and DAP12-/- mice were sorted by flow cytometry. Cells were stimulated with CpG oligonucleotide 2216 for 16 hours. Data are compiled from 6 independent experiments. (B) Mice received intravenous injections of 5 μg CpG 2216 in DOTAP and serum was collected after 6 hours. Representative data of 1 of 3 independent experiments are shown, each with 3 mice per group. IFN-α was measured in cell culture supernatants (A) or mouse serum by ELISA (B). Error bars show one standard deviation. *P < .05 versus B6 by Student t test.
In conclusion, we have identified Siglec-H as a novel murine IPC-specific marker. Siglec-H is unique within the Siglec family, as it associates with the signaling adaptor DAP12, which contains a cytoplasmic ITAM. In contrast, the majority of Siglecs contain cytoplasmic ITIMs. This heterogeneity in the transmembrane and cytoplasmic regions has been previously observed in the NK cell receptor families Ly-49 (in mice) and KIR (in humans),19 and suggests that Siglecs may have evolved from the repeated duplication of an ancestral gene and diverged under selective pressure, perhaps in a sort of arms race between host and pathogens. Siglecs may have developed to recognize self-ligands and mediate tolerance. However, pathogens, such as Neisseria meningitides, have acquired the capacity to encode sialylated molecules that mimic endogenous ligands, in order to attenuate host responses.20 To counteract this pathogen strategy, an originally ITIM-bearing Siglec receptor could plausibly have been converted into a DAP12-associated version capable of providing the host with a pathogen-specific activating receptor. The future identification of the low avidity endogenous glycans and the pathogen-associated glycans recognized by Siglec-H will provide critical insight on the function of this molecule in IPC development and function and in innate immune responses.
Note added in proof. After this article was published online, Zhang and colleagues21 independently described Siglec-H as a marker specific for plasmacytoid DCs.
Prepublished online as Blood First Edition Paper, November 17, 2005; DOI 10.1182/blood-2005-09-3746.
Supported by National Institutes of Health grant R01CA109 673-01A1 (M. Colonna) and Infectious Diseases Training Grant 2T32A10717226 (A.L.B.).
The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 U.S.C. section 1734.
We would like to thank Susan Gilfillan, Julia Klesney-Tait, Isaiah Turnbull, Erik Barton, and Wayne Yokoyama for helpful advice; and William Eades and Jacqueline Hughes in the Siteman Cancer Center High Speed Sorter Core Facility (NCI Cancer Support grant no. P30 CA91842) for cell sorting experiments.