| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Medical Research Council Human Immunology Unit, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Headington, Oxford, United Kingdom;
The Edward Jenner Institute for Vaccine Research, Compton, United Kingdom;
Cochin Hospital, Institut National de la Santé et de la Recherche Medicale, Centre National de la Recherche Scientifique, Université René Descartes, Paris, France; and
Division of Developmental and Clinical Immunology, University of Alabama, Birmingham, AL 35294
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2-microglobulin (2m)-deficient mice develop a multisystem inflammatory disease affecting the joints, skin, and bowel with strong similarity to human spondyloarthritis. We show that HLA-B27 transgenic mice and rats express HC10-reactive, 2m-free HLA-B27 homodimers (B272) and multimers, both intracellularly and at the cell surface of leukocytes, including rat dendritic cells. Fluorescent-labeled tetrameric complexes of HLA-B27 homodimers (B272 tetramers) bind to populations of lymphocytes, monocytes, and dendritic cells. The murine (and probably rat) paired Ig-like receptors (PIRs) are ligands for B272. Thus, B272 tetramers stain RBL cells transfected with murine activating PIR-A4 and inhibitory PIR-B receptors. Murine PIR-A and -B can be immunoprecipitated from the RAW264.7 macrophage cell line, and murine PIR-A can be immunoprecipitated from the J774.A1 line using B272. B272 tetramer staining corresponds to the distribution of PIR expression on lymphoid and myeloid cells and on murine macrophage cell lines. B272 can induce TNF-
release from the J774.A1 macrophage cell line. The binding of B272 to PIR is inhibited by HC10, an mAb that ameliorates arthritis in HLA-B27+ 2m–/– mice. The expression and PIR recognition of B272 could explain the pathogenesis of rodent spondyloarthritis. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
9% of normal controls (2). The pathogenesis of these diseases remains obscure, and the basis for the association with HLA-B27 has not been established.
Animal models of spondyloarthropathy have confirmed the contribution of HLA-B27 to disease pathogenesis. Two transgenic rat lines, 33-3 and 21-4H, carrying high gene copy numbers of HLA-B27 and its human 2-microglobulin (2m)3 partner, consistently develop multiorgan inflammation, resembling human HLA-B27-associated disease (3). Disease features include colitis, enteritis, peripheral and axial arthritis, male genital inflammation, and psoriform skin and nail lesions (3). By contrast, other rat lines (21-4L and 25-6), bearing fewer copies of the transgenes, do not develop disease. HLA-B7 transgenic rats are healthy, and HLA-Cw6 transgenic rats exhibit only transient diarrhea and forepaw swelling (4). Disease can be transferred to HLA-B27 transgenic or nontransgenic rats by transplantation of bone marrow or fetal liver cells from high expressing rat lines (5). Previous studies suggested a role for T cells, with CD4+ cells being more efficient than CD8+ in transferring disease (6). However, recent work has implicated a population of CD8+ monocytes whose numbers are expanded in disease and reduced after ameliorative therapy with an anti-CD8 mAb (7). Disease expression is also dependent on the presence of bacterial flora (8).
Most mice transgenic for HLA-B27 remain healthy, but can develop arthritis and nail changes if the mouse 2m gene is either deleted or replaced with the human homologue (HLA-B27+ 2m–/–, HLA-B27+ 2m–/– h2m+) (9, 10). Symptoms, including arthritis and nail changes, develop on transfer from a specific pathogen-free environment to conventional conditions (9). Although 2m-associated HLA-B27 is not expressed in HLA-B27+ 2m–/– mice, 2m-free HLA-B27 H chains could be detected on the surface of thymic epithelium and Con A-treated lymphoid cells (9, 10). Furthermore, disease development was delayed by treatment with the H chain-specific mAb HC-10 (10). By contrast, an mAb with specificity for 2m-associated HLA-B27 (ME-1) failed to modulate disease (10). This led to the proposal that the expression of free HLA-B27 H chains on the cell surface is involved in the development of spontaneous inflammatory disease in HLA-B27 transgenic mice lacking 2m (10). By contrast, neither murine MHC class II molecules (11) nor TAP-1 (12) were required.
Recently, we have shown that HLA-B27 can form H chain homodimers (B272) in vitro, disulfide-bonded through residue cysteine 67, which are not associated with 2m (13). Furthermore, we have shown that B272 are expressed on the surface of human lymphoid cell lines (13, 14) and in HLA-B27-positive patients with spondyloarthritis (15).
In this study we show firstly that HLA-B27-transgenic animals express HC10-reactive HLA-B27 H chains as homodimers and multimers in a variety of lymphoid cells, both intracellularly and at the cell surface. Secondly, tetrameric complexes of HLA-B27 homodimers (B272 tetramers) bind to populations of myeloid and B cells. The murine (and probably rat) paired Ig-like receptors (PIRs) are ligands for B272, and HC10 mAb inhibits the PIR/B272 interaction.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
HLA-B*2705/human 2m homozygous transgenic mice, obtained from E. Weiss (Institute for Anthropology and Human Genetics, Ludwig Maximilians Universität München, Munich, Germany) (16), were backcrossed onto C57BL/6 to ensure an H2b background. They were then bred with 2m knockout mice (The Jackson Laboratory, Bar Harbor, ME). Mice were bred and maintained at the Edward Jenner Institute of Vaccine Research (Compton, U.K.). Con A-activated blasts were generated from freshly isolated murine splenocytes by culture in DMEM (Invitrogen Life Technologies, Gaithersburg, MD) supplemented with 10% FCS (Invitrogen Life Technologies), 2 mM glutamine, 50 U/ml penicillin, 50 U/ml streptomycin, 50 µM 2-ME, and 20 mM HEPES in six-well plates (2 x 107 cells/well) for 40 h in 8 ml of RPMI 1640 supplemented with FCS, glutamine, HEPES, 2-ME, and Con A (Sigma-Aldrich, St. Louis, MO) to 2.5 µg/ml.
The transgenic rat lines used in this study were bred and maintained at the Cochin Hospital (Paris, France) have been previously described (3) (17) and are detailed in Table I. Rat dendritic cells (DC) were isolated from freshly harvested spleens from transgenic and control rats splenocytes by metrizamide separation after overnight culture as described previously (18) (19).
|
Cell lysis and Western blot analysis
Cell lysis (5 x 106 cells/sample) was conducted in 500 µl of lysis buffer (20 mM Tris (pH 8.0), 150 mM NaCl, 1% Triton X-100, 10 mM iodoacetamide (IAA), and 2 mM PMSF) on ice for 30 min. We have previously shown that postlysis dimerization of HLA-B27 H chains does not occur under these conditions (14). Nuclei and cell debris were pelleted (13,000 rpm 10 min at 4°C). Twenty microliters of lysate was boiled (3 min) in SDS sample buffer (50 mM Tris-HCl, 2% SDS, 10% (v/v) glycerol, and bromophenol blue) with or without the reducing agent DTT (100 mM) and was run on a 10% SDS-PAGE gel. Proteins were transferred onto nitrocellulose membranes (Hybond-C Super; Amersham Biosciences, Piscataway, NJ) using a SemiPhor Unit (Amersham Biosciences). The membranes were blocked overnight at 4°C in blocking buffer (PBS containing 5% (w/v) milk powder (Marvel; Nestlé UK, Croydon, U.K.), and 0.1% Tween 20 (Sigma-Aldrich)). HLA-B27 H chains were detected by incubation with the H chain-specific mAb HC-10 (22) and diluted in blocking buffer, for 1 h at room temperature, followed by washes with blocking buffer, incubation with peroxidase-conjugated goat anti-mouse Ig (Dakocytomation, Carpenteria, CA) and ECL reagent (Amersham Biosciences), and exposure to film (Biomax MR-1; Eastman Kodak, Rochester, NY).
Cell surface labeling and immunoprecipitation
Rat and mouse splenocytes or lymph node cells and DC were surface-labeled using either lactoperoxidase-catalyzed iodination (23) or biotinylation. Briefly, 1–2 x 107 cells/condition were pelleted and washed twice in ice-cold Dulbecco’s PBS. Radioiodination was conducted at room temperature for 15–20 min in 500 µl of iodination buffer (PBS containing 10 µl of lactoperoxidase, 0.2 U/ml glucose oxidase, 5 mM -D-glucose, and 500 µCi of iodine 125) with gentle agitation every 5 min. Unincorporated 125I was removed with washes in PBS/5 mM potassium iodide. Biotinylation of cell surface proteins was performed using EZ-Link Sulfo-NHS-LC-Biotin (Pierce, Rockford, IL). Cell surface proteins were biotinylated by incubating 2 x 107 cells/ml with 0.5 mg/ml EZ-Link Sulfo-NHS-LC-Biotin (Pierce) in PBS at room temperature for 15 min. Cells were then washed once with RPMI 1640 and twice with ice-cold PBS, then lysed for immunoprecipitation. Freshly prepared IAA (10 mM) was included in all steps from cell lysis and in some experiments (e.g., that shown in Fig. 1C) cells were pretreated with IAA for 15 min before biotinylation/iodination. Cell lysates were cleared with rotation at 4°C with 80 µl of protein A-Sepharose beads (10%, v/v; Sigma-Aldrich) for 1 h (14). H chains were then immunoprecipitated with 10 µg/ml HC10 for 1 h, followed by protein A-Sepharose beads (60 µl) for an additional 1 h (4°C). Beads were washed three times with lysis buffer, and a final wash with 10 mM Tris was performed; mAb-Ag complexes were eluted from the beads by boiling (5 min) in SDS-PAGE sample buffer, and supernatants (10–20 µl) were analyzed on 10% SDS-PAGE gels. Gels were fixed in 10% methanol/10% acetic acid and dried down on 3MM paper (Whatman, Clifton, NJ) before autoradiography on Kodak X-OMAT film. Biotinylated blots were developed using streptavidin HRP and ECL reagent (Amersham Biosciences).
|
HLA-B*2705 was expressed and refolded as described previously (13, 15). B272 tetramers were generated in which each HLA-B27 homodimer carried only a single biotin tag, by refolding a mixture of B27-biotin and B27 histidine-tagged proteins in the presence of one of the following known viral peptide epitopes: KRWIIMGLNK (HIV gag), or RRIYDLIEL (EBV EBNA3C). Refolded protein was biotinylated and then purified on Ni-NTA resin (Fast-flow; Amersham Biosciences). Composition was confirmed by nonreducing and reducing SDS-PAGE. PE-labeled ExtrAvidin (Sigma-Aldrich) was then used to generate tetramer complexes.
Standard B27/2m complexes were refolded with 2m and the peptides listed above or with SRYWAIRTR (influenza nucleoprotein 383–391) as previously described (13, 15). Control HLA-A2/2m/flu peptide, HLA-A24 or HLA-B8, or B*702/2m/HIV tetramers were generated under identical conditions as previously described (13, 15). The extracellular domains of HLA-B*702 and B8, expressed under identical conditions and refolded in the same buffer for the same time (48 h) in the presence of cognate peptide, but not 2m, did not form homodimers (S. Kollnberger, A. J. McMichael, and P. Bowness, manuscript in preparation). They did form H chain aggregates, which eluted at the same volume on fast protein liquid chromatography as HLA-B27 aggregates relative to molecular mass standards (corresponding to an Mr of 160,000–250,000). The presence of H chains in these aggregates was verified by analysis by reducing SDS-PAGE. Subsequently, aggregates were biotinylated under identical conditions and used as a control for FACS staining. For the experiment shown in Fig. 4E, biotinylated B272 bound to streptavidin beads was used to immunoprecipitate RAW264.7 lysates (after preclearing with unconjugated streptavidin beads and beads conjugated with HLA-A2/2m/peptide complexes) or J774.A1 lysates (with control immunoprecipitates in parallel). Proteins pulled down by B272 were subsequently resolved by SDS-PAGE and blotted as described above. Blots were developed with the rat anti-PIR mAbs 6C1 or 10.1 (24) (H. Kubagawa, unpublished observations) and HRP-conjugated goat anti-rat Ig (Dakocytomation).
|
|
Twenty-four-well tissue culture plates (Nunc, Naperville, IL) were coated with 0.25 ml of ExtrAvidin (40 µg/ml) in Dulbecco’s PBS, pH 7.4 (endotoxin-free; Sigma-Aldrich), overnight at 4°C. Wells were washed four times with PBS and then blocked with 10% FCS/PBS for 1 h at room temperature. Wells were washed, and 5 µg of biotinylated protein, biotinylated anti-PIRA/B (6C1), or rat IgG1 isotype control mAb (BD Pharmingen, San Diego, CA) was added for 1 h at room temperature. After washing, 0.5 x 106 J774.A1 cells or control M1 cells in RPMI 1640 medium with 10% FCS were added. Polymixin B sulfate (10 µg/ml; Sigma-Aldrich) was used to block LPS-mediated effects. Supernatants were harvested after 24 h for determination of TNF- levels using the murine/rat TNF- ELISA kit (BD Pharmingen) or nitrite levels using Griess reagent (Promega, Madison, WI). Blocking experiments were performed in the presence of 10 µg/ml HC10 or IgG2a isotype control mAb.
FACS staining
Cells (5 x 105 to 1 x 106) were stained on ice with 1–10 µg/ml mAb for 30 min in the presence of azide. In some experiments FcRs were preblocked with 10% normal mouse or rat serum (Sigma-Aldrich) and rat anti-mouse CD16/32-specific mAb 2.4G2 or mouse anti-rat CD32-specific mAb D34-485 (BD Pharmingen). After washing in PBS, 1 mg/ml BSA, and 0.05% sodium azide at 4°C, tetramer staining was conducted at 4°C for 30 min or 37°C for 15 min using 1 µg of tetramer/106 cells in the presence of azide. Cells were resuspended in PBA or were fixed in 2% paraformaldehyde before cytometric analysis on a FACSCalibur using CellQuest software (BD Biosciences, Mountain View, CA). For HC10 blocking experiments, tetramer or cells were preincubated with HC10 mAb or isotype control Ab for 15 min at 4°C before staining as described above.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2m–/– murine splenocytes express HLA-B27 homodimers and multimers
Given that soluble recombinant HLA-B27 H chains can form 2m-free homodimers disulfide-bonded through Cys67 (13), we tested whether they can also dimerize in mouse splenocytes lacking endogenous 2m. Western blot with the mAb HC10, which is specific for free human class I H chains, showed that almost all HLA-B27 H chains are present as dimers with a molecular mass of 90 kDa (Fig. 1A, upper panel, lane 3). Lane 4 shows that dimers were also present in splenocytes transgenic for HLA-B27 and human 2m (HLA-B27+, h2m+, m2m–/–), although a 45-kDa band, almost certainly representing monomeric HLA-B27 H chains associated with h2m, was more abundant. In the presence of DTT, all dimers in the HLA-B27+ 2m–/– mice were reduced to a single 45-kDa monomer band (Fig. 1A, lower panel). The two or three bands seen at 90 kDa under nonreducing conditions probably represent differently disulfide-bonded homodimers. Additional faint, low molecular mass bands were seen after DTT treatment of the HLA-B27+ h2m+ splenocytes, probably representing degradation of HLA-B27 H chains, as observed previously (25).
Con A-stimulated HLA-B27-transgenic 2m–/– mouse splenocytes express low levels of surface HC10 reactivity (9, 10). To determine whether this HC10 staining could be partly due to cell surface B272 expression, Con A-activated splenocytes were surface-labeled with either 125I (Fig. 1B) or biotin (Fig. 1C), and lysates were immunoprecipitated with HC10. Experiments were performed in the presence of IAA to prevent artifactual dimerization. Fig. 1B shows that cell surface HC10-reactive material was composed of HLA-B27 H chain monomers, dimers, and multimers. The right panel shows that these dimers and multimers largely reduced to a 45-kDa band in the presence of DTT. No significant difference in the level of surface HLA-B27 H chain expression, as detected by HC10, was observed between HLA-B27 transgenic and HLA-B27+ h2m+ double-transgenic splenocytes. Surface biotinylation, shown in Fig. 1C, confirmed the presence of HLA-B27 dimer and also higher molecular mass bands. The HC10 Western blot confirmed the presence of HLA-B27 H chains in the dimer band reducing to 45-kDa monomers and showed that the 80-kDa band seen on the reduced gel was not HLA-B27 (data not shown).
In HLA-B27 transgenic rats, splenocytes, lymph node cells, and DC express HLA-B27 H chains in homodimeric and multimeric forms
We next studied lysates from lymph node and spleen mononuclear cells from the HLA-B27 transgenic and nontransgenic rat lines shown in Table I. Cells from the diseased 33-3 HLA-B27 rat with a high copy number of the HLA-B27 transgene showed high expression levels of B272 (Fig. 2A, lane 4). Three prominent B272 bands of 90 kDa were resolved. B272 accounted for 50% of the total H chains by densitometry (data not shown). Their levels were lower in 21-4L rats with fewer copies of HLA-B27 and were lower still (<10% total H chains) in rats transgenic for HLA-B7, which lacks Cys67. Almost all dimers were thiol-labile and reduced to a single 45-kDa band in the presence of DTT (Fig. 2A, right panels).
|
2m. Tetrameric complexes of HLA-B27 H chain homodimers stain populations of murine B cells and monocytes
To test whether B272 bind to murine cells, fluorescent B272 tetramers were generated. Fig. 3, A and C, shows that B272 tetramers consistently stained 1–10% of splenocytes within the lymphocyte gate. These were CD3 negative (Fig. 3A), CD19 positive (Fig. 3, B and C) B lymphocytes; their frequencies did not differ significantly between HLA-B27 transgenic and nontransgenic mice. These cells scarcely bound standard HLA-B27 and HLA-A2 tetramers above levels observed with ExtrAvidin-PE. Binding of the B272 tetramers to B lymphocytes could be blocked by HC10 mAb, but not by IgG2a isotype control mAb (Fig. 3C).
|
|
B272 tetramers bind murine PIRs; binding is inhibited by HC10 mAb
We noted that the same mouse and rat cell populations that bind B272 tetramers also express PIRs (e.g., murine B, myeloid, and DC lineages, but not T or NK cells) (24, 26). In repeated experiments similar numbers of murine B lymphocytes and monocytes were stained with B272 tetramer and with the PIR-specific 6C1 mAb (Fig. 4A). 6C1 is a nonblocking mAb that binds both PIR-A and PIR-B (24). This correlation extended to murine monocyte/macrophage cell lines expressing PIRs (Fig. 4B). Thus, homodimer tetramers failed to bind to the monocytic cell line M1 that showed the weakest PIR expression. B272 tetramer staining of the J774.A1 cell line was inhibited with the HC10 mAb (Fig. 4C).
We next studied RBL-2H3 cells transfected with murine PIR-A4 or PIR-B. Fig. 4D shows that B272 tetramers specifically bind to PIR-expressing transfectants; moreover, staining correlated with the level of PIR expression. By contrast, neither control HLA-A2 or HLA-B27 heterodimers tetramers nor PE-conjugated B*0702 and B8 H chain aggregate bound to the RBL-PIR transfectants. PE-conjugated HLA-B27 H chain aggregates stained PIRA4-transfected RBL-2H3 cells (Fig. 4D, second experiment). Weaker staining of RBL-PIRB transfectants was also observed (data not shown). The B272 tetramer staining could be partially inhibited by preincubating the tetramers with the HC10 mAb or by preincubating the cells with a 10-fold excess of unlabeled B272 monomer (data not shown).
We next used B272-coated beads to immunoprecipitate lysates from the RAW264.7 and J774.A1 monocytic cell lines. Proteins were eluted from beads, resolved by 12% SDS-PAGE, and subjected to Western blot with the rat anti-PIR mAbs 6C1 or 10.1. Fig. 4E (left panel) shows that B272 clearly immunoprecipitates two PIR-reactive bands of 85 and 120 kDa, corresponding to PIR-A and -B, respectively. The band around 22 kDa represents material at the dye front. Fig. 4E (right panel) shows that B272 (but not heterodimeric complexes) immunoprecipitates PIR-A from J774.A1 cells.
B272 tetramers, but not HLA-A2 or HLA-B27 heterodimer tetramers, stain populations of rat B cells, CD4+ CD8+ monocytes, and DC
We next used B272 tetramers to stain splenocytes and lymph nodes from healthy and transgenic rat strains. B272 tetramers did not stain rat T lymphocytes (Fig. 5A, left panel), but stained 2–8% of B cells (Fig. 5A, right panel). B272 tetramers bound B cells from HLA-B27 transgenic, HLA-B7 transgenic, and nontransgenic rats (data not shown).
|
B272 induces TNF release from J774.A1 cells
Fig. 6 shows a representative experiment in which B272 complexes immobilized on streptavidin-coated plates specifically induced TNF- release by the J774.A1 macrophage cell line. ExtrAvidin alone or standard HLA-A2 heterodimeric complexes gave only a low/background level of TNF production (Fig. 6A). Production of TNF- by J774.A1 cells could also be stimulated by cross-linking FcRs in the presence of IFN-
. Similar results were obtained when B272 complexes were preincubated with cells before addition of extravidin (data not shown). All experiments were set up and performed in the presence of 10 µg/ml polymyxin B sulfate, which abrogated LPS-induced TNF production (see Fig. 6C), but had no significant effect on B272-induced TNF production (data not shown). The lower panel of Fig. 6A shows that plate-bound anti-PIRA/B mAb (6C1), but not isotype control rat IgG1, also induced TNF- secretion by J774.A1 cells. Fig. 6B shows that B272 also induced nitrite release from J774.A1 cells. Nitrite release was increased in the presence of the cross-linking anti-FcR Ab 2.4G2; this release was inhibited by HC10, but not by an isotype IgG2a mAb. Fig. 6C shows that B272 complexes did not induce TNF- from the M1 cell line, which expresses little PIR and does not stain significantly with B272 (see Fig. 4B).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2m–/– mice (10), inhibits the B272/PIR interaction. HLA-B27-transgenic mice and rats express B27 homodimers and multimers
HLA-B27 H chain expression has been previously detected with the HC10 mAb in HLA-B27+ m2m–/– mice (9, 10). We show that a substantial proportion of these H chains are expressed as 2m-free homodimers and multimers, both intracellularly and on the surface of murine splenocytes and rat DC. We propose that these molecules have a role in disease pathogenesis, either through intracellular expression of B272 and induction of a proinflammatory stress response (27) or, more likely, though cell surface recognition by the immune system. Intracellular B272, resolved by SDS-PAGE under nonreducing conditions, show three bands at 90 kDa. This molecular mass heterogeneity is also seen in HLA-B27-transfected human cell lines, where cysteine mutagenesis suggests that differing intra- or intermolecular disulphide bond pairing is responsible (14). Two-dimensional isoelectric focusing under reducing and nonreducing conditions confirms the presence of B272 (data not shown) (14, 15). We have previously shown in cell mixing experiments that B27 dimer formation does not occur postlysis under the conditions used in this study (15). The HC10mAb also immunoprecipitated surface-biotinylated HLA-B27 H chains after IAA pretreatment (Fig. 1C), making it unlikely that cell surface dimers are an artifact of the iodination process. The additional band(s) seen on HC10 IP of surface-labeled material would be consistent with the presence of additional molecules, such as ERp57 (25) or BIP (GRP78) (28), disulphide-bonded to HLA-B27 H chains.
H chain homodimer formation has also been observed previously for certain murine alleles, apparently via their unpaired cytoplasmic Cys (23). This could also explain the low levels of a single 90-kDa band seen in this study for HLA-B7, which lacks the unpaired Cys67 in the 1 helix (14). Homodimerization (and homotrimerization) through unpaired cysteine residues has also recently been described for the nonclassical HLA-G (29). We propose that HLA-B27 homodimerization stabilized by Cys67 could lead to a unique homodimer conformation. This is supported by our data showing that cell surface expression of B272 in human cells is dependent on the unpaired Cys67 (14). It is also possible that the higher molecular mass multimeric forms of HLA-B27 may play a significant functional role in vivo.
Studies with murine MHC class I molecules have suggested that H chain dimer formation is related to the availability of 2m (23). In its absence, dimers form in the ER and can traffic to the cell surface in murine cells (as confirmed in this study, shown in Fig. 1B); in its presence, they can form in a post-Golgi compartment after 2m dissociation from unstable complexes. The recent finding that on certain backgrounds disease can occur in 2m–/– mice without any need for the HLA-B27 transgene (28) suggests that dimerization of some mouse alleles might contribute to spondyloarthropathy, and that 2m could be protective. Indeed, it is possible that murine free H chains or H chain dimers are the natural ligands for PIRs (see below).
The dependence on high HLA-B27 copy number for rat disease (17) is also consistent with a role for excess H chains and suggests a quantitative effect, perhaps through saturation of peptide supply or other Ag presentation machinery. Notably under certain conditions other alleles can be induced to form homodimers in vivo (28).
Murine PIRS bind B272: rat PIR expression could account for the distribution of homodimer receptors
We have shown that B272 bind to cells of the monocyte/macrophage lineage and to a subset of B cells from HLA-B27-transgenic and nontransgenic rodents of different strains. B272 tetramer staining of murine cells ex vivo correlates with PIR expression on B cells and monocytes, but not T or NK cells (24). B272 tetramers stained PIR-transfected RBL cells and PIR-expressing macrophage cell lines. Moreover, PIR-A and -B could be immunoprecipitated from macrophage cell lines by B272, but not HLA-A2 or HLA-B27 heterodimers. It is possible that some of the other homodimeric or H chain structures besides B272 may bind receptors such as PIRs. We have attempted to control for this by generating PE-conjugated B*0702 and B8 H chain aggregate. These did not bind to PIR-A transfected cells. By contrast, PE-conjugated H chain aggregates of HLA-B27 were able to bind PIR transfectants. Binding of B272 to PIRs could be inhibited with the HC10 mAb, which has previously been reported to ameliorate disease in HLA-B27-transgenic 2m knockout mice (10). B272 complexes induced TNF and nitrite production by the PIR-expressing murine macrophage line J774.A1.
In the mouse there are at least six Pira genes, encoding (putative) activatory PIR-A receptors with short cytoplasmic tails and a charged arginine residue in their transmembrane domain that facilitates association with adaptor proteins such as the FcR common -chain to form a cell activation complex (30). PIR-B is encoded by a single Pirb gene (31) and contains four potential ITIM motifs in the cytoplasmic tail (32). We show that both the murine activatory PIR-A4 and the inhibitory PIR-B are ligands for B272.
The rodent PIRs share considerable sequence homology (40–60%) with the human leukocyte Ig-like receptor (LILR)/leukocyte Ig receptor (LIR) family of receptors (26). We have also shown that standard heterodimeric complexes of HLA-B27, -B8, or -A2 with 2m and antigenic peptides are not ligands for murine PIRs. Although the murine ligands for PIRs remain to be elucidated, it has recently been shown that the human MHC class Ib molecule HLA-G can bind to the murine PIR-B (33). Binding of HLA-G tetramers to bone marrow-derived DCs induced phosphorylation of PIR-B receptors and inhibited cellular maturation. In keeping with this observation, HLA-G transgenic mice have compromised maturation of functional DCs (33), as do HLA-B27 transgenic rats (19). Taken together these results raise the possibility that PIRs might be receptors for murine free class 1 H chains or for nonclassical MHC molecules. It is possible that interaction of B272 with one of the activating PIR-A isoforms may drive inflammation through enhanced cytokine production (as shown in this study for TNF-). To this end it may be significant that rat B-lymphocytes express only activating PIR-A receptors. In contrast myelomonocytic cells express PIR-B as well as PIR-A, and it is possible that interaction of B272 with PIR may inhibit or modulate immune function. Notably PIR-B–/– mice show impaired DC maturation and increased Th2 responses (34). Interestingly, LPS results in significant up-regulation of PIR expression on B cells and macrophages (24), suggesting one possible explanation for the necessity of bacterial flora for disease generation (8, 9).
Although our results suggest a model in which B272 expressed by APC interact with PIRs on monocytes or B cells to induce or perpetuate immunopathology, our findings do not exclude a role for NK cells (which in the rat can express the inhibitory PIR-B (26)) or T lymphocytes through either indirect mechanisms or induction of PIR expression under distinct conditions. It is also important to point out that B272-reactive NK family receptors can be expressed by human NK and T cells (see below), and that rodent receptors with sequence homology to human NK receptors have recently been identified (35).
Rat DC express B272, B27 multimers, and receptors for B272
We have shown that splenic DCs from HLA-B27 transgenic rats with inflammatory disease express significant levels of cell surface B27 homodimers and multimers. Previous cell transfer experiments have shown that transplantation of bone marrow or fetal liver cells from high expressing HLA-B27 transgenic rats can elicit disease in both transgenic and nontransgenic recipients (5). It is, therefore, possible that the expression of B272 by bone marrow-derived DCs is involved in disease pathogenesis. We have also shown that rat DC express receptors for B272. These are likely to be PIR; rat myeloid cells express both PIR-A and -B. Coexpression of both B272 and B272 ligand could have significant functional implications and might explain the impaired DC function described in HLA-B27 transgenic rats (19).
Implications for human spondyloarthritis
Patients with ankylosing spondylitis express B272 on mononuclear cells (15) and show increased levels of HC10 staining on monocytes (36). Moreover, there are receptors for B272 in these patients (15). The observations in humans and rodents are consistent with a role for the interaction of B272 with immunoreceptors in inflammatory disease. Crucial to this hypothesis, we have shown firstly that B272 receptors are present in rodent models of disease, and secondly that B272 binds a different set of immunoreceptors to normally conformed HLA-B27/2m/peptide molecules in both rodents and humans. Thus, B272 tetramers bind to the human immunoreceptors LILRB2 (also known as LIR2 or ILT4), (37), LILRA1 (37), killer Ig-related receptor 3DL1 (KIR3DL1) (37), and KIR3DL2 (15), but not to LILRB1 (LIR1 or ILT2) (37). In contrast, HLA-B27/2m/peptide complexes do not bind to PIRs significantly and bind human LIR1, but not KIR3DL2 (37) (15). There are, however, species differences in the cellular expression pattern of B272 receptors between humans and rodents. Whereas B272 tetramers did not stain murine or rat CD3+ T lymphocytes, we have observed binding to up to 6% of human T cells (in addition to B cell and monocyte staining), consistent with binding to three domain KIR receptors (15).
Evidence for a key role of TNF- in spondyloarthritis
We have shown that B272 can induce TNF- release from macrophages in vitro. TNF- expression has been demonstrated within the inflamed sacroiliac joints of patients with ankylosing spondylitis (38), where macrophages and T cells are demonstrable (39). Furthermore, recent therapeutic trials with anti-TNF- blockers in human disease have shown striking clinical benefit (40, 41) and radiological resolution of inflammation on magnetic resonance image scanning (42). Such benefits have not been reported with T cell-targeted therapies.
Although our data suggest a pathogenic role for cell surface B272 through interaction with immunoreceptors, other mechanisms may be involved. Mear and colleagues (27) have suggested that intracellular expression of B272 may induce a proinflammatory stress response. We have shown that B272 are abundantly expressed intracellularly in rodent disease models and could indeed set up an unfolded protein stress response. Our data are also compatible with models in which B272/PIR interactions might costimulate the generation of peptide-specific T cell responses, commensurate with data suggesting a role for CD4+ T cells in rat disease (6) and the finding of CD4 and CD8 T cell responses to triggering bacteria in human reactive spondyloarthritis (43, 44, 45).
In conclusion, our data show that HLA-B27 H chains are expressed as dimers and multimers in transgenic rodent models of spondyloarthritis. We also show that both mice and rats express receptors for B272 on B cells and cells of the monocyte/macrophage lineage. PIRs are the principal B272 receptor in mice and probably also in rats. We provide evidence that B272 can have a functional effect on murine macrophages. These findings suggest that the interaction of B272 with immunoreceptors on cells of the myelomonocytic lineage or on B lymphocytes might be involved in the pathogenesis of spondyloarthritis.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 This work was supported by the Medical Research Council of the United Kingdom and the Arthritis Research Campaign. 
2 Address correspondence and reprint requests to Dr. Paul Bowness, Medical Research Council Human Immunology Unit, University of Oxford, Oxford, U.K. OX3 9DS. E-mail address: pbowness{at}gwmail.jr2.ox.ac.uk
3 Abbreviations used in this paper; 2m, 2-microglobulin; B272, HLA-B27 H chain homodimer; DC, dendritic cell; h, human; IAA, iodoacetamide; KIR, killer Ig-related receptor; LILR, leukocyte Ig-like receptor; LIR, leukocyte Ig receptor; PIR, paired Ig-like receptor.
Received for publication November 3, 2003. Accepted for publication May 17, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2m: an animal model of HLA-B27-associated human disorders. Cell 63:1099.[Medline]
T cells are not essential to the pathogenesis of arthritis or colitis in HLA-B27 transgenic rats. J. Immunol. 170:1099.2-microglobulin: a model of human spondyloarthropathies. J. Exp. Med. 182:1153.2-microglobulin (2m) double transgenic mice with disrupted mouse 2m. J. Clin. Invest. 98:2746.[Medline]
2-microglobulin-free heavy chain homodimer structure. J. Immunol. 162:5045.2-microglobulin dissociation. J. Immunol. 151:159.[Abstract]
2-microglobulin-deficient mice without expression of HLA-B27: association with deficiency of endogenous major histocompatibility complex class I expression. Arthritis Rheum. 43:2290.[Medline]
chain. J. Exp. Med. 188:991.. N. Engl. J. Med. 346:1349.This article has been cited by other articles:
|
|
 |
|
  S. Endo, Y. Sakamoto, E. Kobayashi, A. Nakamura, and T. Takai Regulation of cytotoxic T lymphocyte triggering by PIR-B on dendritic cells PNAS, September 23, 2008; 105(38): 14515 - 14520. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Masuda, A. Nakamura, T. Maeda, Y. Sakamoto, and T. Takai Cis binding between inhibitory receptors and MHC class I can regulate mast cell activation J. Exp. Med., April 16, 2007; 204(4): 907 - 920. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Shiroishi, K. Kuroki, L. Rasubala, K. Tsumoto, I. Kumagai, E. Kurimoto, K. Kato, D. Kohda, and K. Maenaka Structural basis for recognition of the nonclassical MHC molecule HLA-G by the leukocyte Ig-like receptor B2 (LILRB2/LIR2/ILT4/CD85d) PNAS, October 31, 2006; 103(44): 16412 - 16417. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Turner, D. P. Sowders, M. L. DeLay, R. Mohapatra, S. Bai, J. A. Smith, J. R. Brandewie, J. D. Taurog, and R. A. Colbert HLA-B27 Misfolding in Transgenic Rats Is Associated with Activation of the Unfolded Protein Response J. Immunol., August 15, 2005; 175(4): 2438 - 2448. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
