| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia;
MacFarlane Burnet Institute of Medical Research and Public Health, Austin Campus, Heidelberg, Victoria, Australia; and
Kirin Brewery, Tokyo, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
(Sirp1 and Sirp4) molecules were identified as differentially expressed in CD8– cDC. Genomic sequence analysis revealed a third Sirp member localized in the same gene cluster. These Sirp genes encode cell surface molecules containing extracellular Ig domains and short intracytoplasmic domains that have a charged amino acid in the transmembrane region which can potentially interact with ITAM-bearing molecules to mediate signaling. Indeed, we demonstrated interactions between Sirp1 and 2 with the ITAM-bearing signaling molecule Dap12. Real-time PCR analysis showed that all three Sirp genes were expressed by CD8– cDC, but not by CD8+ cDC or plasmacytoid pre-DC. The related Sirp
gene showed a similar expression profile on cDC subtypes but was also expressed by plasmacytoid pre-DC. The differential expression of Sirp and Sirp1 molecules on DC was confirmed by staining with mAbs, including a new mAb recognizing Sirp1. Cross-linking of Sirp1 on DC resulted in a reduction in phagocytosis of Leishmania major parasites, but did not affect phagocytosis of latex beads, perhaps indicating that the regulation of phagocytosis by Sirp1 is a ligand-dependent interaction. Thus, we postulate that the differential expression of these molecules may confer the ability to regulate the phagocytosis of particular ligands to CD8– cDC. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
DC can be broadly classified into conventional DC (cDC) (2) and plasmacytoid pre-DC (pDC) (3). pDC are found in all lymphoid organs and can be distinguished from cDC on the basis of morphology, surface phenotype, and the ability to secrete high levels of IFN-. These pDC only assume a typical DC morphology and function after activation. The cDC found in lymphoid organs can be further separated based on surface phenotype. In the thymus, there are at least two DC populations that can be distinguished on the basis of CD8 expression (CD8– and CD8+) (4). These thymic cDC are thought to be critical in the process of thymocyte-negative selection and thus in central immune tolerance. In the spleen, there are three subpopulations that can be distinguished based on expression of CD4 and CD8, the double-negative CD4–8– (DN), the CD4+8– (CD4+), and the CD4–8+ (CD8+) (4). In lymph nodes (LN), three additional migratory populations have been identified, including the DC derived from skin epidermal Langerhans cells (LC), from dermal DC, and from epithelial DC draining the lung (5, 6). These spleen and lymph node DC are thought to be critical in regulating the balance between effective responses to invading organisms and maintenance of peripheral tolerance to self-Ags.
Different DC subtypes display important differences in specialized immune functions, including their production of cytokines and chemokines, the type of T cell response they initiate, and their processing of Ags for MHC I or MHC II presentation (recently reviewed in Ref. 2). The molecules that differ between these DC subtypes are therefore of interest, because they may underpin these functional differences. Surface molecules differing between the DC subtypes are of special interest, because they may serve as targets for selective delivery of Ags or therapeutic agents to modulate immune responses. In previous studies, we used representational difference analysis to identify two novel cell surface molecules, Fire (Emr-4) and Cire (CD209a), expressed on the surface of CD8– but not CD8+ cDC (7, 8). Here, we continue to investigate murine DC gene expression using custom DC cDNA microarrays constructed from a subtracted, normalized pan-DC gene expression library. By comparing gene-expression profiles of the three mouse spleen cDC subtypes, we have identified then characterized a family of signal regulatory proteins (Sirp) as novel, differentially expressed, DC surface components.
Sirps are surface molecules containing extracellular Ig-like domains and are classified in human as SIRP, SIRP, or SIRP
molecules. SIRP (CD172a, Shps-1, p84, BIT, and MyD1) is expressed in neurons, myeloid cells, and DC, and has a long intracytoplasmic tail (110–113 aa) containing two potential ITIM-signaling motifs. Human SIRP (SIRP1), expressed in macrophages and DC, has a short intracytoplasmic domain (5 aa) that lacks signaling motifs but has a charged amino acid residue in the transmembrane domain (reviewed in Refs. 9 and 10) which interacts with ITAM-bearing molecules (11, 12, 13). Human SIRP (SIRP2), which is expressed in T cells, B cells, and NK cells, lacks both the cytoplasmic ITIMs that are characteristic of SIRP and the transmembrane lysine that is characteristic of SIRP1 (14, 15, 16). SIRP, and to a lesser degree, SIRP, binds CD47 on the surface of interacting cells (15, 16, 17, 18, 19). In SIRP, this results in signaling through the ITIMs and is implicated in macrophage and DC functions such as cell motility, phagocytosis, and cytokine production (20, 21, 22, 23, 24). In the case of SIRP, the interaction with CD47 appears to be adhesion based and is implicated in the enhancement of T cell proliferation (15). No ligand has been identified for Sirp to date and little is known about its function, although recent reports suggest an activatory role in macrophage phagocytosis (25). We now characterize three murine Sirp genes encoding DC surface molecules, and show that Sirp and Sirp are differentially expressed by CD8– cDC.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
All mice were bred under specific pathogen-free conditions at the Walter and Eliza Hall Institute (WEHI) animal breeding facility. For all experiments, female C57BL/6J WEHI mice of 6–9 wk of age were used.
Isolation of DC
DC isolations were performed from lymphoid organs as previously described (3, 4). To generate activated DC, mice were injected i.p. with 20 nM fully phosphorothioated CpG (Geneworks) according to a published sequence (26) and 1 µg of LPS from Escherichia coli (Sigma-Aldrich), and the DC were isolated 3 or 16 h later. The DC-enriched preparations were stained with fluorochrome-conjugated Abs and flow cytometric sorting performed using a MoFlo instrument (DakoCytomation) with autofluorescent cells excluded. cDC were sorted for high forward scatter and high expression of CD11c and further segregated based on CD4 and CD8 expression for splenic DC, CD205 and CD8 expression for lymph node cDC, and CD8 expression for thymic cDC. pDC were flow cytometrically sorted based on intermediate expression of CD11c and high expression of CD45RA.
Sirp protein surface expression
To examine surface expression of Sirp proteins on hemopoietic cells, cells were prepared by flow cytometric sorting as follows: T and B cells were isolated by preparing spleen and LN cell suspensions, removing erythrocytes and dead cells then incubating with fluorochrome-conjugated Abs to distinguish the B cells (B220+CD3–) and the T cells (B220–CD3+). Macrophages were prepared from spleen and bone marrow cell suspensions by removing erythrocytes and dead cells, and the splenic macrophages (Mac-1+CD3–B220–) and the bone marrow macrophages (Mac-1+GR-1–B220–) distinguished. Splenic NK cells were prepared by collagenase digestion of the spleen, magnetic bead depletion of cells using anti-CD3, anti-CD90, anti-TER119, anti-GR-1, and anti-CD19, before distinguishing the NK1.1+CD49b+ splenic NK cells. Thymocytes were prepared from a thymus cell suspension and stained with anti-CD3, anti-CD4, and anti-CD8. All DC or cell suspensions were blocked using rat Ig and anti-FcR mAb (2.4G2) before immunofluorescence staining. The mAb to Sirp proteins, the p84 Ab to Sirp (donated by Dr. C. Lagenaur, University of Pittsburgh, Pittsburgh, PA) and the 39/04-7E11 Ab to Sirp1 (described below), were biotinylated and streptavidin-PE or streptavidin-Alexa 594 were used as detection reagents. Control staining using biotinylated isotype control mAbs (BD Pharmingen), rat IgG1 for the p84 anti-Sirp and rat IgG2a for the anti-Sirp1, was performed. Fluorescent labeled cells were analyzed on a FACStar Plus instrument (BD Biosciences). A positive control sample (cDC) was stained and analyzed in each experiment to control for differences between experiments in reagents and instrument settings.
Generation of DC microarrays
Total cDC and pDC were isolated from the thymus, spleen, and LN of steady-state mice and from mice injected i.p. with CpG and LPS. Organs were harvested 16 h later and total RNA isolated using an RNeasy Mini kit (Qiagen). The total RNA from steady-state and activated DC from the different tissues was pooled and poly(A)+ RNA prepared using an Oligotex mRNA kit (Qiagen). Amplified DC cDNA was generated using a SMART amplification system (BD Clontech). To generate a cDNA library that was enriched for DC-specific genes, a PCR-select cDNA subtraction system (BD Clontech) was used to allow subtractive hybridization against cDNA purified from cultured NIH 3T3 fibroblast cells and normalization by suppression PCR. The efficacy of the subtraction and normalization procedures was validated using semiquantitative PCR, to check for representation of two DC genes, Fire and Cire, known to be expressed at low levels (7, 8) and two housekeeping genes, Gapdh and -actin. In the subtracted DC cDNA, the weakly expressed DC genes Cire and Fire were represented 30- to 1000-fold more often than the housekeeping genes, in contrast to the original DC RNA, where they were expressed at 0.03-fold the level of these housekeeping genes (data not shown), indicating the subtraction and normalization of the DC library was successful. The normalized and subtracted DC cDNA was subcloned into a pGemT easy vector (Promega) and transformed into XL10 Gold competent cells (Stratagene). Approximately 10,000 clones were picked at random, the inserts PCR amplified using the primers (5'-tcgagcggccgcccgggcaggt-3'; 5'-agcgtggtcgcggccgaggt-3'), purified using Montage PCR96 plates (Millipore) and printed using 50% DMSO onto amino-silane Corning GAPSII slides using a Virtek printer.
Microarray analysis
Amplified cDNA was generated from 1 µg of RNA by SMART amplification (BD Clontech), labeled with Cy3-dCTP or Cy5-dCTP (Amersham Biosciences) by random primed labeling (Invitrogen Life Technologies) and purified using an Autoseq G-50 column (Amersham Biosciences). The Cy-dye labeled amplified cDNA was then hybridized to the custom microarray slides for 16 h (27) and the slides scanned using a GenePix 4000B scanner (Axon Instruments).
To make comparisons with the best possible accuracy, the microarray experiments were conducted as direct designs, also referred to as saturated or loop designs (28). Each microarray directly compared samples from two DC subtypes, rather than using a common reference RNA source (see Fig. 1B). Two dye-swapped biological replicates were used for each comparison to balance any possible dye effects for each gene and to allow meaningful statistical analysis. Spot intensities were acquired using the image analysis program SPOT (CSIRO Australia, 
http://experimental.act.cmis.csiro.au/Spot
) using "morph" background correction. Log ratios were print-tip loess normalized (29). Linear modeling methodology was used to extract full information on each comparison from the complete experiment (30, 31). Empirical Bayes moderated t-statistic methods were used to ensure stable statistical results even with modest numbers of replicate arrays (31). As each clone was spotted twice on each microarray, correlation methods were used to maximize the information from these within-array replicate spots (32). The high quality of the arrays in conjunction with the above bioinformatic treatment mean that it was not necessary to filter low intensity spots, thus preserving all information on less highly expressed genes. All differentially expressed genes reported in Table I have p < 2 x 10–8. Statistical analysis used the Limma software package for the R programming environment (http://bioinf.wehi.edu.au/limma).
|
|
To determine the sequence of the differentially expressed clones in the microarray analysis, an aliquot of the corresponding frozen bacterial clone was cultured and Miniprep plasmid DNA prepared (Qiagen). Sequencing was performed using the Big Dye Terminator version 3.1 (Applied Biosystems) with M13(–20) primer (gtaaaacgacggccagt) and 200 ng of plasmid DNA, and subjected to electrophoresis on an ABI 3730xl 96-capillary automated DNA sequencer. Comparison of sequences to the expressed sequence tag, cDNA and protein databases was performed by basic local alignment search tool alignment using National Center for Biotechnology Information (www.ncbi.nlm.nih.gov). Genomic localization was performed by BLAT alignment to the mouse assembly (May 2004) from the University of California, Santa Cruz, CA, Genome Bioinformatics (www.genome.ucsc.edu/).
Quantitative RT-PCR
RNA (up to 1 µg) was DNase treated with RQ1 DNase (Promega) then reverse transcribed into cDNA using random primers (Promega) and Superscript II reverse transcriptase (Invitrogen Life Technologies). Real-time RT-PCR was performed to determine the expression of Sirp, Sirp2, Sirp1, Sirp4, and Gapdh in splenic DC subsets using the Quantitect SYBR Green PCR kit (Qiagen) and a Light cycler (Roche). The specific primers for real-time RT-PCR were as follows: Sirp; 5'-accaccgtgaaccctagtggaa-3'; 5'-ggtgggtgaaactcggatgaag-3', Sirp2-f1,2: 5'-gcggcactgagttgtttgtcct-3'; 5'-gtcctacacctcggtaccacctc-3', Sirp1: 5'-ggattctggtcccaacagggttctta-3'; 5'-atatgcacagctttaaagttctggga-3', Sirp4: 5'-ctgccgacattgtcccatttttct-3'; 5'-tccacagaaccctgctcatgactctac, Gapdh; 5'-catttgcagtggcaaagtggag-3'; 5'-gtctcgctcctggaagatggtg-3'. An initial activation step for 15 min at 95°C was followed by 40 cycles of: 15 s at 94°C (denaturation), 20–30 s at 50–60°C (annealing), and 10–12 s at 72°C (extension), followed by melting point analysis. The expression level for each gene was determined using a standard curve prepared from 10–2–10–6 pg of specific DNA fragment, and was expressed as a ratio relative to Gapdh.
Expression of Sirp proteins
Full-length cDNA was isolated for the Sirp genes by PCR amplification from splenic DC cDNA using Advantage cDNA polymerase (BD Clontech) and the resultant products subcloned into pGemT easy plasmid (Promega). Sirp proteins were expressed on the surface of Chinese hamster ovary (CHO) cells as N-terminal (extracellular) FLAG-tagged proteins and on the surface of rat basophilic leukemia (RBL) cells as fusion proteins where GFP was fused to the C-terminal cytoplasmic domain of Sirp. To generate the FLAG-tagged proteins, Sirp cDNA was amplified using Pwo polymerase (Roche), restriction digested with Mlu-1 and subcloned into a pEF-Bos vector modified to contain the IL-3 leader sequence followed by the FLAG epitope (donated by Dr T. Willson, WEHI). CHO cells were cotransfected with the pEF-Bos-Sirp and a pCI-neo plasmid containing the neomycin phosphotransferase gene (Promega) using FuGENE 6 (Boehringer Mannheim) and transfectants selected with 1 mg/ml G418 (Geneticin; Invitrogen Life Technologies). Sirp-positive cells were stained with biotinylated anti-FLAG mAb M2 (Sigma-Aldrich), followed by streptavidin-FITC (BD Pharmingen), and then isolated by flow cytometric sorting. GFP-tagged proteins were generated by amplifying Sirp cDNA, restriction digesting with HindIII and XmaI, and subcloning into pEGFP-N1 vector (BD Clontech), before electroporation (Gene Pulsar; Bio-Rad) into RBL cells and selection with 500 µg/ml G418. Sirp-positive cells were isolated by flow cytometric sorting of GFP-positive cells.
Generation of mAb against Sirp proteins
Wistar rats were immunized three to four times with 1.5 x 107 CHO cells expressing Sirp-FLAG at 4-wk intervals, and given a final boost 4 days before fusion. Hybridomas secreting specific mAb were identified by flow cytometric analysis of supernatants using RBL cells expressing GFP-Sirp. One hybridoma 39/04-7E11 secreting a rat IgG2a Ab to Sirp1 was identified that showed reactivity with GFP-Sirp on the surface of RBL cells and Sirp-FLAG on the surface of CHO cells, but showed no reactivity with parental RBL or CHO cells. It also showed no cross-reactivity with Sirp2 and Sirp4 (data not shown).
Phagocytosis of opsonized RBC by macrophages
Bone marrow-derived macrophages (BMM) were prepared by culturing C57BL/6 mouse bone marrow in 10% L-cell conditioned DMEM for 7 days. A total of 105 BMM were seeded per sterile coverslip in a 24-well tray and allowed to adhere overnight at 37°C in a 10% CO2 incubator. The cells were incubated with 20 µg/ml anti-Sirp Ab (39/04-7E11) or a rat IgG2a isotype control (GL117), followed by 20 µg/ml goat anti-rat IgG (Jackson ImmunoResearch Laboratories). Opsonized mouse RBC were generated by incubation of the RBC with 2 µg/ml rabbit anti-mouse RBC IgG (Cedarlane Laboratories), before incubating them with the BMM at a ratio of 10:1 for 15 min at 37°C. Unphagocytosed RBC were lysed, the BMM were fixed with methanol and stained with Giemsa, the percentage of BMM that had phagocytosed RBC was enumerated by light microscopy, and the results were expressed as mean ± the SEM.
Phagocytosis of amastigotes by DC
Total splenic cDC (5 x 106) were incubated with 20 µg/ml anti-Sirp Ab (39/04-7E11) or a rat IgG2a isotype control (GL117), followed by 20 µg/ml goat anti-rat IgG Ab (Jackson ImmunoResearch Laboratories) as described above. Leishmania major amastigotes were isolated from skin lesions and draining lymph nodes of infected CBA/H nu/nu mice (33) and used to infect DC at a 20:1 ratio for 2 h at 37°C. The DC were then allowed to adhere for 2h at 37°C on coverslips precoated with anti-MHC class II Abs. The cells were washed and fixed with methanol and stained with Giemsa, as previously described (33). The percentage of infected DC was calculated after counting 500 cells on each duplicate slide and the results were expressed as mean ± the SEM.
Immunoprecipitations
CHOP cells (CHO cells stably transfected with polyoma T Ag) were transiently cotransfected by lipofection with FLAG-tagged mouse Sirp1 and Sirp2 cDNA expression constructs, in the presence or absence of MYC-tagged mouse Dap12 (34) (a gift from Drs. M. Smythe and P. Darcy, Peter MacCallum Cancer Institute, Melbourne, Australia). A construct expressing FLAG-tagged membrane protein Fire (7) was included in these analyses as a control. After 2 days, 1 x 107 cells were lysed in 1% Triton X-100, 150 nM NaCl, 1 mM EDTA, 10 mM Tris-HCl (pH 7.4) in the presence of protease inhibitors. Lysates were precleared using an irrelevant mouse IgG Ab and protein G-Sepharose beads (Sigma-Aldrich). FLAG-tagged proteins were immunoprecipitated with mouse anti-FLAG (M2; Sigma-Aldrich), and washed four times with lysis buffer. Immune complexes were separated by electrophoresis using a 10% SDS-PAGE gel under nonreducing conditions and immunoblotted with mouse anti-c-MYC (9E10; Sigma-Aldrich) followed by rat-anti-mouse IgG-HRP (BD Biosciences). HRP-labeled proteins were visualized by Supersignal West Pico chemiluminescence (Pierce).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
A custom DC cDNA library enriched for DC-specific genes was generated using a PCR-select cDNA subtraction system (BD Clontech) with subtractive hybridization of pooled DC cDNA against cultured NIH 3T3 fibroblast cDNA. The NIH 3T3 cell line was chosen as it is distinct from lymphoid and myeloid cells, and should thus allow subtraction of genes expressed in all cell types, with no expected loss of DC-specific genes. Normalization of the cDNA library by suppression PCR was included to increase the relative frequency of weakly expressed genes.
Comparison of gene expression patterns between splenic DC subsets
To compare the gene expression profiles between the splenic cDC subsets, we purified cDC from mouse spleen and sorted the CD4–8– (DN), the CD4+8– (CD4+) and the CD4–8+ (CD8+) cDC subsets (Fig. 1A). cDNA amplification was performed from 1 µg of RNA (27) due to the scarcity of sorted mouse DC and a direct microarray comparison performed between each of the splenic subsets (Fig. 1B). Microarray analysis indicated that there were differences in gene expression profiles between all three splenic cDC subsets (Fig. 1C). However, in a comparison of the DN and the CD4+ cDC, although there were numerous spots representing gene fragments differentially expressed by the DN cDC, there were very few spots representing gene fragments that were expressed by the CD4+ but not the DN cDC (Fig. 1C). Sequence analysis of differentially expressed gene fragments from our microarray analysis identified a range of molecules, including genes already known to be differentially expressed on DC subsets such as CD1d1 (35), CD4 (4), CD8 (4), CD86 (4), CD209a/Cire (8), and Emr-1/F4/80 (4), as well as some genes that were novel or previously uncharacterized in DC (Table I). The chemokines MIP-1/CCL3, MIP-1/CCL4 and RANTES/CCL5 were also identified as preferentially expressed by the CD4+ DC, and we have recently reported the differential expression of chemokines by DC subtypes (36).
Of interest were genes with sequence similarity to the SIRP family of surface molecules that were differentially expressed in the CD8– cDC, (that is in the DN and the CD4+ subsets) (Table I). We identified 14 clones covering three fragments of a gene Sirp (GenBank accession NM_198405.1) and 1 clone of a gene fragment encoding a novel sequence with some similarity to known Sirp molecules, referred to hereafter as Sirp1 and Sirp4, respectively. Both Sirp1 and Sirp4 appeared preferentially expressed in the CD8– cDC. Both Sirp1 and Sirp4 showed high sequence similarity to human SIRP molecules and encoded Sirp molecules. That is, they contained extracellular Ig domains, short intracytoplasmic domains, and a conserved lysine residue in the transmembrane domain, which in human SIRP is required for interaction with an aspartic acid residue of the adaptor protein DAP12.
Analysis of the public databases, including the expressed sequence tag and genomic databases, suggested that Sirp1 and Sirp4 were two genes in a cluster of three Sirp-like genes, localized on mouse chromosome 3. Hence, we designed primers to investigate the expression of Sirp1, Sirp4, and the third gene, Sirp2 (GenBank accession XM_355437.1), by quantitative RT-PCR. Interestingly, Sirp (Shps-1, p84, BIT), the mouse ortholog of human SIRP, has previously been described to be differentially expressed by mouse CD8– splenic DC, based on representational difference analysis (I. Caminschi and M. D. Wright, unpublished observations) and microarray gene profiling (35) and to be differentially expressed by rat splenic DC (37). Therefore, we investigated the expression of each of the Sirp genes in mouse splenic cDC subsets and confirmed that all four Sirp genes were more highly expressed in the CD8– cDC relative to the CD8+ cDC (Fig. 1D).
Identification, characterization, and cloning of the Sirp gene family
We amplified the full-length cDNA encoding each of the Sirp molecules by PCR and sequenced the genes. Each of the Sirp genes on chromosome 3 contains a single open reading frame that encodes putative transmembrane proteins with Ig domains in their extracellular regions, a short cytoplasmic tail, and a transmembrane region containing a lysine residue, similar to human SIRP1. The Sirp genes found in humans, rats, and mice appear to have diverged from a common ancestral Sirp gene after species differentiation and consequently these genes do not have obvious orthologs.
The full-length coding sequence of Sirp2 is encoded by four exons spanning at least 58.3 kb of genomic DNA. Full-length Sirp1, which was also cloned by Hayashi et al. (25), is encoded by five exons spanning over 45.9 kb. Full-length Sirp4 is encoded by three exons spanning at least 35.8 kb of chromosome 3 (Fig. 2A). The predicted protein structure of the Sirp molecules is schematically represented in Fig. 2B, demonstrating the transmembrane structure and the extracellular Ig domains of the Sirp family.
|
2 open reading frame (ORF) (900 bp, submitted to GenBank, accession DQ055451) encodes a protein of 299 aa that has two V-type Ig domains in the extracellular region (Fig. 3A). The Sirp1 ORF (1176 bp, submitted to GenBank, DQ055453) encodes a 391-aa protein that has one V-type Ig domain and two C1-type Ig domains (Fig. 3B). The Sirp4 ORF (537 bp, submitted to GenBank, accession DQ055452) encodes a protein of 178 aa with a single V-type Ig domain (Fig. 3C). Although the Sirp genes differ in the number and type of Ig domains in their extracellular regions, they all show a high level of sequence identity in their leader sequences, in their V-type Ig domains (Fig. 3D), and in their transmembrane and cytoplasmic domains.
|
genes contained a charged transmembrane lysine residue that could potentially mediate interactions with ITAM-bearing adaptor molecules such as Dap12. To test this possibility, FLAG-tagged Sirp1, and Sirp2 expression constructs, were transiently transfected into CHOP cells, in the presence or absence of the adaptor signaling molecule Dap12 (tagged with the MYC epitope). Transfectants were lysed, putative complexes were isolated by immunoprecipitation, and associated Dap12 was detected by Western blot analysis. Molecular interactions between Sirp1 or Sirp2, with Dap12 could readily be detected (Fig. 2C), and were not detected between Dap12 and irrelevant FLAG-tagged membrane proteins (data not shown). Expression of Sirp genes
We examined the expression of the Sirp genes across a panel of organs and cell types by quantitative real-time RT-PCR. Although the Sirp gene expression profiles were similar, there were some significant differences between the Sirp and the Sirp gene expression profiles. Sirp was expressed mainly in bone marrow, spleen, LN, lung, and brain whereas the Sirp genes showed highest expression in bone marrow, spleen, and lung (Fig. 4A). In the hemopoietic compartment, both Sirp and Sirp were expressed by NK cells, mature macrophages, and DC. Sirp2 was also expressed highly in immature macrophages (Fig. 4B).
|
and Sirp genes in DC subsets, the highest levels of Sirp and Sirp gene expression were observed in the DN and CD4+ splenic cDC and in the CD8– lymph node cDC. Sirp was also expressed in splenic and thymic pDC, whereas Sirp genes were not expressed in pDC (Fig. 5A). Upon in vivo activation with CpG and LPS, ligands to TLRs 9 and 4, respectively, the levels of Sirp gene expression were substantially reduced in both the DN and the CD4+ splenic cDC, whereas the levels of Sirp genes were not affected (Fig. 5B). Interestingly, each of the three Sirp genes showed similar expression profiles across a range of cell and tissue types although the actual levels of mRNA transcripts varied between the three genes.
|
and Sirp1
To investigate the surface expression of Sirp1, we generated a mAb (39/04-7E11) that demonstrated flow cytometric reactivity to Sirp1, but not to Sirp2 or Sirp4 on the surface of transfectant cells (data not shown). The surface expression of Sirp and Sirp1 on cDC and pDC was then investigated by staining splenic and thymic DC with mAb to Sirp or to Sirp1 followed by flow cytometric analysis (Fig. 6). Sirp1 expression was observed on splenic cDC (nonautofluorescent CD11chighCD45RA–), but was completely absent from splenic pDC (nonautofluorescent CD11cintCD45RA+), or thymic cDC and pDC. In contrast, high levels of Sirp were observed on the majority of splenic cDC, whereas thymic cDC segregated into a major population of Sirplow and a minor population of Sirphigh cells. Intermediate levels of Sirp were observed on splenic and thymic pDC, although the levels observed on splenic pDC were higher than the thymic pDC. These surface expression profiles were consistent with the gene expression levels of Sirp and Sirp1 observed by quantitative RT-PCR.
|
was also expressed at intermediate levels on most splenic and bone marrow macrophages, but only marginally on splenic NK cells. However, Sirp1 was expressed at intermediate levels on approximately half the splenic macrophages and at low levels on a very small proportion of the bone marrow macrophages (Fig. 7). Sirp and Sirp1 expression was undetectable in splenic (Fig. 7) and lymph node T and B cells and in thymocytes (data not shown). Thus, in the hemopoietic system, both Sirp and Sirp1 can be considered to be essentially restricted to expression on DC and macrophages, albeit with differences in the levels of expressions (Fig. 7).
|
and Sirp1 were then compared on splenic, LN, and thymic cDC. The cDC were gated for CD11chigh nonautofluorescent cells, and then segregated on the basis of CD4 and CD8 expression for splenic cDC, on CD205 and CD8 expression for LN cDC, and on CD8 expression for thymic DC (Fig. 8A). Both Sirp and Sirp1 were differentially expressed by the DC subsets (Fig. 8, B and C).
|
and were negative for Sirp1. The different CD8– cDC subtypes varied in their level of expression of both Sirp and Sirp1. Sirp expression was generally high on all CD8– cDC subtypes, including the minor CD8– subset of thymic cDC. However the migratory forms of cDC in s.c. LN showed heterogeneity in expression, with some expressing comparatively low levels, especially among the dermal-derived DC. This may have reflected an increased level of activation or maturation within these migratory DC, because Sirp mRNA levels declined with deliberate DC activation, as shown previously. Sirp1 staining was always lower than Sirp staining. Sirp1 expression was clearly positive, if relatively low, on the splenic DN and CD4+ cDC subsets, but was low to negative on all other CD8– cDC in LN and spleen, and largely negative on the migratory DC populations in LN.
Use of Sirp as a marker to discriminate DC subtypes
It was clear that Sirp could serve as a useful marker to aid in the discrimination of mouse DC subtypes, particularly in combination with CD8 (Fig. 8D). In spleen, there was a clear separation of CD8–Sirphigh and CD8+Sirplow subgroups. In thymus, the minor subgroups of CD8– DC stood out as Sirp+, a particularly useful segregation because separation by CD8 staining alone is complicated by the pickup of CD8 from thymocytes (4). In LN, the presence of additional migratory DC subsets produces a more complex situation, but Sirp staining may help resolve the several CD8– DC subtypes.
Role of Sirp1 in phagocytosis
To examine the biological effects of ligating the Sirp1 molecules using the mAb 39/04-7E11, we initially investigated the ability of BMMs to phagocytose, as Sirp and more recently Sirp1 have been implicated in macrophage phagocytosis of IgG-opsonized RBC (21, 25). BMM were incubated with the anti-Sirp1 Ab or a rat IgG2a isotype control with or without the addition of an anti-rat IgG antiserum. Use of the anti-Sirp1 Ab with a secondary Ab resulted in an inhibition of RBC phagocytosis, whereas no inhibition was observed in the absence of the secondary Ab, or with the isotype control (Fig. 9A). Mouse splenic DC do not phagocytose opsonized RBC, hence we investigated the role of Sirp1 ligation in phagocytosis of known DC targets, such as FITC-labeled latex beads and L. major amastigotes. Use of the anti-Sirp1 with a secondary Ab resulted in an inhibition of L. major amastigote phagocytosis (Fig. 9B), whereas no inhibition was observed in the absence of the secondary Ab or in the presence of the isotype control Ab. Furthermore, consistent with the expression profile of Sirp1, preliminary studies using sorted splenic cDC demonstrated that the inhibition of phagocytosis by the anti-Sirp1 Ab was specific to the DN and CD4+ cDC and was not observed in the CD8+ cDC (data not shown). In contrast, the anti-Sirp1 Ab, with or without the secondary Ab, had no effect on the phagocytosis of FITC-labeled latex beads (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Microarray comparison of the splenic cDC subtypes revealed differences in gene expression profiles between the CD8+ and CD4+ subsets, and between the CD8+ and DN subsets. Interestingly, in a comparison of the CD4+ and the DN, although there were numerous genes that were differentially expressed by the DN and not the CD4+, there were very few that were expressed by the CD4+ and not the DN cDC. These data are consistent with studies by Edwards et al. (35) of gene expression profiles of splenic DC using commercial arrays. This suggests that the CD8+ cDC subset is most divergent compared with the DN and the CD4+ subsets, which appear to be more closely related. This is certainly supported by functional studies of splenic subsets in which the CD8+ cDC are distinguished by a selective capacity to cross-present exogenous Ags on MHC class I, having the greatest capacity to produce IL-12p70 and thus induce Th1 responses, a reduced capacity to secrete chemokines (36), a restricted capacity to initiate proliferation and cytokine production by T cells compared with CD8– cDC (2), and distinct TLR expression (38, 39). The DN and CD4+ cDC subsets are more closely related functionally, although it should be noted that they are not directly precursor-product related (40). One possible reason for the biased difference in gene expression is that the splenic DN population may be heterogeneous, with the majority of cells being closely related to the CD4+ cDC and a minor proportion of cells being more divergent, as previously suggested by others (35).
The analyses of gene expression using our custom microarrays led to the identification of a family of three Sirp surface molecule encoding genes localized in a cluster on mouse chromosome 3. The Sirp molecules are all differentially expressed by quiescent CD8– cDC subsets of the spleen and lymph nodes (Figs. 5 and 8). The ITIM-bearing molecule Sirp showed a similar expression profile with higher levels of expression in the CD8– cDC of the spleen and lymph nodes. However, there were significant differences in the pattern of Sirp and Sirp expression. Unlike Sirp, Sirp is strongly expressed in pDC. It is also differentially expressed in thymic DC where the two DC subsets are convincingly separated into CD8+Sirplow and CD8–Sirphigh populations. Finally, Sirp gene expression is down-regulated upon activation of DC with LPS and CpG, possibly indicating an inhibitory role for this molecule in immature DC. The differential expression of these molecules in DC subsets may have important functional implications. Sirp binds/interacts with its ligand CD47 on the surface of cells such as T lymphocytes, erythrocytes, and endothelial cells to mediate its downstream functions. Sirp regulates cytoskeletal reorganization and cell motility (20). It has been shown to regulate the migration of Langerhans cells, monocytes, and neutrophils (41, 42, 43). It has also been shown to play an important role in regulating phagocytosis of RBC and platelets by macrophages (21, 23), and apoptotic cells by phagocytes (macrophages and immature DC) (22). Sirp has also been implicated in cytokine production by human monocytes and DC, and in the maturation of human DC (24, 44). Furthermore, SIRP/CD47 interactions between human DC and T cells suppress IL-12 production in the former and IL-12 responsiveness in the latter cell type (24). This observation correlates with differential expression of Sirp on the CD8– populations in mouse, as these populations are poor IL-12 producers and are thought to promote Th2 immunity (2).
In contrast, relatively little is known about the function of Sirp molecules. These molecules contain between 1 and 3 extracellular Ig domains and short intracytoplasmic domains. They demonstrate high sequence identity in their extracellular IgV, transmembrane, and cytoplasmic domains. The extracellular Ig domains are likely to be required for protein-protein interaction with other cell surface molecules. Despite its high similarity to human SIRP at the extracellular domains, human SIRP1 is unable to bind CD47 and no ligands have been reported for SIRP1 or any of the mouse Sirp molecules to date. The similarity of the mouse Sirp molecules and human SIRP1 in the transmembrane and cytoplasmic region and the conservation of the lysine residue in the transmembrane domain, which in human and mouse Sirp1 is required for interaction with ITAM-containing molecules such as DAP12, suggests the three Sirp molecules signal through a similar mechanism through ITAM-bearing molecules. Indeed, we could readily demonstrate a molecular interaction between Dap12 with Sirp1 or Sirp2 in a transient transfection system (Fig. 2C).
In macrophages, Sirp1 engagement, using the mAb generated in our laboratory, inhibited phagocytosis of opsonized RBC (Fig. 9) and L. major promastigotes and amastigotes (data not shown), The Ab also inhibited phagocytosis of L. major amastigotes by DC (Fig. 9). It is of some interest that these effects were only observed after the addition of a secondary goat anti-rat Ab, which might suggest that cross-linking Sirp1 at the cell surface initiates a signal transduction cascade that inhibits phagocytosis. However, the observation that phagocytosis of FITC-coupled latex beads was not affected by Sirp1 engagement, argues strongly that the role of Sirp1 in phagocytosis is dependent on interactions with specific ligands on the opsonized RBC or Leishmania surface, interactions that are effectively blocked by Sirp1 mAb when complexed with secondary Ab. Sirp (Sirp1) engagement has been previously reported to play a role in cytoskeletal reorganization and phagocytosis in macrophages (25). Surprisingly, Sirp1 engagement using anti-Sirp1 mAbs induced macrophage phagocytosis in the studies of Hayashi et al. (25), whereas our mAb inhibited phagocytosis. Thus, while it is clear that Sirp1 plays a role in phagocytosis, whether it induces or inhibits its biological effect/function may be dependent on the epitope bound by the different Abs. The CD8– cDC subsets are significantly more effective than their CD8+ counterparts at phagocytosis be it Leishmania (33, 45) or latex beads (40). Thus, the expression of Sirp and Sirp molecules may play an important role in regulating phagocytosis of particular ligands in CD8– cDC subtypes.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by grants from the National Health and Medical Research Council of Australia and Kirin Brewery. 
2 Address correspondence and reprint requests to Dr. Mireille H. Lahoud, The Walter and Eliza Hall Institute of Medical Research, 1G, Royal Parade, Parkville, Victoria 3050, Australia. E-mail address: lahoud{at}wehi.edu.au
3 Abbreviations used in this paper: DC, dendritic cell; cDC, conventional DC; pDC, plasmacytoid pre-DC; DN, double negative; Sirp, signal regulatory protein; LN, lymph node; LC, Langerhans cell; RBL, rat basophilic leukemia; BMM, bone marrow-derived macrophage; ORF, open reading frame.
Received for publication July 19, 2005. Accepted for publication April 7, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
with KARAP/DAP-12. Eur. J. Immunol. 30: 2147-2156. [Medline]1 is a DAP12-associated activating receptor expressed in myeloid cells. J. Immunol. 164: 9-12. 2-CD47 interaction costimulates T-cell proliferation. Blood 105: 2421-2427. [Medline]1. Eur. J. Immunol. 30: 2130-2137. [Medline]
