| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
: A Potential Role in Systemic Lupus Erythematosus1
* Department of Medicine and Department of Immunology, University of Colorado Health Sciences Center, Denver, CO 80262;
Mary Kirkland Center for Lupus Research, Hospital for Special Surgery, New York, NY 10021;
Randall Centre for Molecular Mechanisms of Cell Function, King’s College London, New Hunt’s House, Guy’s Campus, London, United Kingdom;
Department of Biochemistry and Molecular Genetics, University of Colorado Health Sciences Center, Denver, CO 80262;
¶ Department of Molecular and Computational Biology, University of Southern California, Los Angeles, CA 90089; and
|| Laboratory of Molecular Biology, National Institute of Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD 20892
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
. IFN-, a multifunctional cytokine important in the innate immune system, has recently been proposed to play a major pathogenic role in the development of systemic lupus erythematosus in humans and mice. In this study, we have shown using surface plasmon resonance and ELISA approaches that CR2 will bind IFN- in the same affinity range as the other three well-characterized ligands studied in parallel. In addition, we show that IFN- interacts with short consensus repeat domains 1 and 2 in a region that serves as the ligand binding site for C3d/iC3b, EBV-gp350, and CD23. Finally, we show that treatment of purified human peripheral blood B cells with the inhibitory anti-CR2 mAb 171 diminishes the induction of IFN--responsive genes. Thus, IFN- represents a fourth class of extracellular ligands for CR2 and interacts with the same domain as the other three ligands. Defining the role of CR2 as compared with the well-characterized type 1 IFN- receptor 1 and 2 in mediating innate immune and autoimmune roles of this cytokine should provide additional insights into the biologic roles of this interaction. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
CR2 has been primarily viewed as a B cell coreceptor whose biologic role is to amplify Ag receptor-mediated signal transduction. In support of this conclusion, coligation of CR2 with the membrane BCR results in enhanced intracellular calcium release, proliferation, and/or activation of MAPK, as well as up-regulation of B7 molecules (11, 12, 13, 14, 15). Enhancement of BCR-dependent cell activation by CR2 is due to the association of CR2 with CD19 and CD81 in a B cell-specific signal transduction complex (16, 17), where coligation of BCR with CR2/CD19 amplifies signals without efficient engagement of inhibitory phosphatases (14, 15).
In addition to enhancing BCR-dependent signals, ligation of CR2 alone has been reported to result in several other phenotypes, including the induction of homotypic adhesion (18), NF-
B activation (19), IL-6 generation (20), Ag uptake and presentation to T cells (21, 22), and actin polymerization (23).
CR2 has been shown to be one of several receptors for human CD23 (3, 4). CD23 is an immunoregulatory protein found both on cell membranes and as a soluble protein. CD23 interacts with CR2 to increase production of IgE in the presence of IL-4 (3), rescue germinal-center B cells from apoptosis (24), provide T cell-activating signals by B cell APCs (25), and promote T-B cell adhesion (26).
In mice, CR2 is encoded along with the larger receptor CR1 by the Cr2 gene, which produces both proteins through alternative splicing of a common mRNA (27). Support for a critical role of CR2 in the immune response has been provided by results in Cr2–/– mice that are deficient in CR2 (as well as mouse CR1). Cr2–/– mice demonstrate substantial defects in Ag-specific, T cell-dependent and T cell-independent humoral immune responses (28, 29, 30) that is due to a lack of receptor on both B cells and follicular dendritic cells (31). Cr2–/– mice also demonstrate defects in B cell memory (32, 33) and the development of the natural Ab repertoire (34, 35).
In addition to having an important role in the immune response to foreign Ags, recent studies have strongly suggested that CR2 plays a key role in maintaining tolerance to self-Ags such as ssDNA, dsDNA, chromatin, and histones, as well as the development of experimental systemic lupus erythematosus (SLE) (13, 36, 37). The molecular explanation for this phenotype is incompletely understood. One hypothesis suggests that enhanced autoimmunity to nuclear Ags in the relative or complete absence of CR2 and/or CR1 function may be due to the requirement for complement and CR2/CR1 in Ag capture in the bone marrow (36, 38). Another hypothesis is that, in the absence of CR2/CR1, there is ineffective deletion of autoreactive B cells by C3/C4-bound Ags (13, 36). Although the exact role of CR2 in maintaining tolerance to DNA, histones, and chromatin is unclear, markedly lower expression of human CR2 in patients with SLE (39) and mouse models of SLE before the onset of detectable humoral autoimmunity (40) have suggested that alteration of this function of CR2 is likely to be pathophysiologically important in the development of human SLE.
IFN- has recently been proposed as a molecule that is also centrally implicated in the pathogenesis of SLE, a hypothesis that is supported by several lines of investigation. For example, a subset of patients with SLE demonstrate a peripheral blood lymphocyte mRNA profile consistent with IFN- stimulation (41). In addition, lupus-prone New Zealand Black mice whose expression of the type I IFN- receptor (IFNAR) has been eliminated by gene targeting demonstrate amelioration of autoimmune disease (42).
One previous series of experiments has shown that IFN- blocks EBV binding to, and capping of, CR2 on B cells (43, 44). In addition, IFN- was reported to contain a peptide sequence similar to one from C3d that was proposed to be at the C3d-CR2 binding interface (43). That particular IFN--derived peptide was reported to bind to purified CR2 in vitro and moderately blocked the interaction of iC3b/C3d-coated zymosan particles with B cells. Although these data provided some evidence that IFN- could be an authentic ligand for CR2, the conclusions from this study were challenged by the finding that the C3d-derived peptide sequence used as the basis for the IFN- comparison was found to not be near the CR2 binding site for C3d in the C3d-CR2 cocrystal structure (6).
For this reason, in addition to the emerging interests related to the potential roles of both IFN- and CR2 in the development of SLE, we have re-examined the interactions of this ligand-receptor pair. To provide a relative context for this ligand-receptor study, we have also compared the affinity and kinetics of IFN- binding to CR2 with three other well-characterized ligands for this receptor. We find, using surface plasmon resonance, ELISA, and IFN--responsive gene analysis, that IFN- exhibits readily detectable, high-affinity interactions with CR2. These data suggest that the roles of these two proteins in the development of autoimmunity may be mechanistically linked in the pathogenesis of SLE.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Human CR2 SCR1-4 was amplified by PCR from a previously used CR2-encoding expression vector (45) using the primers (5'-CCG GGC CAG CCG GGC CAT TTC TTG TGG CTC TCC TCC GCC-3' and 5'-GAA TGC GGC CGC ACT ACT AAA AAT TTC TTC ACA TAC TGG CAT TTT GG-3') that also generate Sfi and NotI restriction endonuclease enzyme sites at the 5' and 3' ends, respectively. A clone containing aas 70–156 of the biotin carboxyl carrier protein (BCCP) was provided by Dr. J. Lambris (Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, PA), as was a clone encoding the biotin holoenzyme synthetase protein Bir A, which catalyzes the incorporation of biotin into BCCP (46). A Bir A enzyme kit was also purchased from Avidity. Sequences encoding CR2 SCR1-4 alone, or with BCCP, were then cloned into the pSecTag2/Hygro vector (Invitrogen Life Technologies) such that soluble SCR1-4 would be expressed with and without C-terminal BCCP (Fig. 1). Before transfection, each cDNA underwent nucleotide sequence analysis (University of Colorado Cancer Center) to assure the presence of the anticipated DNA sequence.
|
The cDNAs encoding the recombinant proteins SCR1-4 and SCR1-4-BCCP were stably expressed in the NSO and HEK 293 mammalian cell lines using electroporation and Lipofectamine 2000 (Invitrogen Life Technologies), respectively. For NSO cell line expression, the CR2 SCR1-4 cDNA was linearized and then electroporated at 100 microfarads/300 V in 0.5 ml of a PBS solution containing 50 µg of herring sperm DNA and 6 µg of SCR1-4 cDNA. Cells were then serially diluted in medium under selection with 800 µ/ml hygromycin. For HEK 293 cell expression, cells were recovered by 0.1% trypsin treatment and release from a flat-bottom flask, counted, and then seeded at 2.5 x 105 in 500 µl growth medium containing no serum and no antibiotics. For transfection of the CR2 SCR1-4-BCCP cDNA, 1 µg of linearized plasmid and 3 µl of Lipofectamine 2000 in 50 µl of Opti-MEM (Invitrogen Life Technologies)-reduced serum medium was mixed and incubated at 37°C for 30 min, followed by incubation in tissue culture wells for 24–48 h. Following this incubation, cells were diluted serially and grown in the presence of 800 µg/ml hygromycin. In both expression systems, after 2–3 wk, individual clones were recovered and screened for CR2 protein expression by ELISA. The secreted protein obtained from both of these cell line supernatants was concentrated and stored at –70°C until sufficient volume was ready for purification.
CR2 SCR1-4 and CR2 SCR1-4-BCCP were both purified in an identical fashion. Supernatant was concentrated and dialyzed in 50 mM PBS overnight at 4°C. Dialyzed supernatant was loaded on a previously equilibrated HB-5-Sepharose 4B anti-CR2 mAb (which specifically recognizes an epitope in CR2 SCR3-4) affinity column at 4°C. Following washing of unbound protein, bound protein was eluted using a buffer previously described (47) to maintain CR2 activity that contains 0.2 M acetic acid, 150 mM NaCl, 1 mM PMSF, 1 µM leupeptin, and 1 µM pepstatin (pH 2.5). The eluate was immediately neutralized with 1.0 M Tris buffer (pH 7.4). The eluted protein was pooled, concentrated, and dialyzed in 50 mM PBS overnight at 4°C. All CR2 proteins purified through this method were then finally passed through a protein G-Sepharose column to remove a trace amount of anti-CR2 mAb HB-5 that leaches from the column during this purification method.
CR2 SCR1-2 and CR2 SCR1-15 expression and purification
CR2 SCR1-2 produced in Pichia pastoris and CR2 SCR1-15 produced in baculoviral Sf9 cells were produced as described by Guthridge et al. (48). Briefly, the human CR2 SCR1-2 domain was cloned into the expression plasmid pPICZ (obtained from Invitrogen Life Technologies) and used to transfect the X33 wild-type strain of P. pastoris. A highly producing transfected clone was grown by fermentation. To purify CR2 SCR1-2, the cell supernatant was buffered to pH 4.0, and the supernatant was passed over a SP-Sepharose column and eluted using 10 mM formate at pH 4.0 with a 0–0.5 M linear NaCl salt gradient. The eluted CR2 SCR1-2 material was pooled, concentrated, and deglycosylated overnight at 37°C with 33,000 U EndoH/ml (New England Biolabs). The deglycosylated sample was passed over SP-Sepharose, and a 0.0–0.5 M NaCl gradient was used to elute the deglycosylated CR2 SCR1-2. Positive fractions were pooled and concentrated. The protein size and purity was assured by both nonreducing and reducing SDS-PAGE analysis using NuPAGE 10% Bis-Tris gels and MES NuPAGE running buffer (Invitrogen Life Technologies). Aminoterminal amino acid sequencing and MALDI-TOF mass spectroscopy analysis was also performed to confirm the exact molecular mass. This same material was also used in the cocrystalization studies with C3d (6).
Human CR2 SCR1-15 was expressed in Sf9 insect cells as a recombinant soluble protein in the baculovirus expression system. The cDNA coding for the first 15 SCR domains plus the native human Kozac consensus sequence and signal peptide as a soluble protein was cloned into the pVL1393 baculoviral expression vector (BD Pharmingen). After infection of the insect cells with the recombinant baculovirus, spinner cultures were grown for 5 days at 27°C. The supernatants were removed, filtered, and concentrated, and the concentrate was dialyzed overnight into 50 mM PBS at 4°C. To purify CR2 SCR1-15, the dialyzed supernatant was loaded on a previously equilibrated HB-5 Sepharose mAb column, and protein was purified as noted above for CR2 SCR1-4.
Human C3dg-biotin and C3d expression and purification
Human C3dg-biotin (49) and C3d (48) were grown and purified as described previously. Briefly, C3dg-biotin was cloned into the pET11b vector (Novagen). This construct consisted of a bicistronic bacterial expression vector with a T7 promoter, an N-terminal polyhistidine, and a T7 epitope tag, followed by the human C3dg sequence (C3 residues 953 to 1303, in which Cys1010 at its thioester site was mutated to Ala: Swiss-Prot code P01024), a GGGSGGGS linker, and a C-terminal biotinylation signal peptide (BSP; LGGIFEAMKELRD) for birA-catalyzed biotinylation. A second ribosomal binding site allows for translation of the birA gene in vivo in cells that also express the recombinant C3dg-BSP protein. C3dg-biotin was produced in the Escherichia coli BL21 pLysS Codon Plus strain (Stratagene) transfected with the C3dg bicistronic vector. The cultures were induced with 25 mM isopropyl 
-D-thiogalactoside (IPTG) at 28°C and shaken overnight. The harvested pellets were resuspended in PBS at pH 8.0 with Complete EDTA-free protease inhibitor tablets (Roche-Boehringer Mannheim) and lysed using four freeze thaw cycles. DNase (4,000 U) and RNase (50 µl of 0.5 mg/ml stock) were added, and the lysate was clarified by sonication and centrifugation at 8,000 x g, followed by mixing for 30 min with 10 ml of Talon (Co2+-immobilized metal affinity chromatography) resin. Unbound material was washed off using PBS at pH 8.0 and 10 mM imidazole, and the C3dg-biotin was eluted by a linear gradient to 100 mM imidazole and further purified using a Superdex-75 gel filtration column (Amersham Biosciences). Biotinylation was confirmed by Western blots using streptavidin (SA)-HRP, which detected bands that also reacted with a rabbit anti-C3dg Ab.
Human C3d was expressed by transfection of the pET15b-C3d plasmid into BL21 pLysS Codon Plus E. coli (Stratagene). Cultures were grown as described for C3dg-biotin above, the bacterial pellet was harvested in DEAE starting buffer (20 mM Tris; pH 7.1) plus Complete EDTA-free protease inhibitor tablets, and lysis was performed as described above. C3d was purified using a DEAE HiPrep column, followed by a MonoQ HR5 column in 20 mM Tris (pH 7.1) and a MonoS HR5 column in 50 mM MES (pH 7.0), where in all three cases C3d was eluted using a linear salt gradient up to 1 M NaCl. The protein purity was confirmed using SDS-PAGE.
EBV gp350 expression and purification
As previously described (50), a 70-kDa fragment of the EBV gp350/220 coat protein was cloned into the pVI-Bac transfer vector (BD Pharmingen). This contained a melittin signal sequence and a recombinant protein with a C-terminal polyhistidine tag. This fragment of gp350/220 was produced by infecting Sf9 insect cells with the gp350/220-packaged baculovirus particles. The cultures were grown, and the culture supernatant was processed as described above for CR2 SCR1-15. The 70-kDa fragment of gp350/220 was then purified using a Con A Sepharose column. The concentrated supernatant was loaded on a previously equilibrated column. The unbound protein was washed off with 20 mM Tris-HCl (pH 7.4) containing 0.5 M NaCl. The bound protein was eluted with increasing concentration of -D-methylmannoside or -D-methylglucoside. The eluted protein was concentrated, dialyzed into PBS, and stored at –20°C. Protein purity was assessed by SDS-PAGE.
CD23 expression and purification
Using the numbering from Swiss-Prot accession no. P06734, derCD23 comprises the amino acids Ser156 to Glu298. The human derCD23 construct was subcloned from CD23 cDNA by PCR into pET5a. Overexpression of CD23 was induced with 0.4 mM IPTG in the E. coli host strain, BL21 (DE3) (pLysS) (Invitrogen Life Technologies) cultured in M9 minimal medium. The cells were incubated for a further 3 h at 37°C then harvested and stored at –20°C. The recombinant protein was present as inclusion bodies and extracted as reported (51), including 100 U of DNase I in the first solution, and replacing Nonidet P-40 with Igepal (Sigma-Aldrich). The final inclusion body preparation in 6 M guanidine was refolded via a glutathione intermediate (52). CD23 was purified by hydrophobic interaction chromatography using Phenyl Sepharose 6 (high substituted) Matrix (Amersham Biosciences) equilibrated with 25 mM Tris-HCl (pH 7.5), containing 1.5 M ammonium sulfate. The refolded protein was adjusted to 1.5 M ammonium sulfate and filtered through a cellulose acetate filter (0.8-µm pore; Sattorius). Purified CD23 was eluted by a decreasing ammonium sulfate concentration gradient. Fractions containing CD23 were pooled, concentrated by ultrafiltration (10-kDa cutoff YM10 cellulose membranes; Millipore), and dialyzed into 25 mM Tris-HCl (pH 7.5), 137 mM NaCl, and 2 mM CaCl2.
IFN- source for biophysical studies
Human recombinant human (rh)IFN- in phosphate buffer was obtained from PBL Biomedical Laboratories. For surface plasmon analysis, exchange of IFN- into the working buffer was accomplished using a Centricon (Millipore).
Bir A enzyme expression and purification
The biotin holoenzyme synthetase protein, Bir A, catalyzes the incorporation of biotin into BCCP. This protein was expressed in E. coli using a clone that was provided by Dr. J. Lambris (Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, PA) and purified according to previously described method (46). Following growth at 37°C in YT medium containing 50 µg/ml ampicillin, 0.25 mM IPTG was added to induce protein expression. The expressed protein in the cell lysate was purified by Ni-NTA affinity column. Briefly, the cell lysate was dialyzed in 10 mM Tris, 300 mM NaCl, and 10 mM imidazole (pH 8.0) overnight at 4°C. The dialyzed supernatant was incubated with 5 ml of Ni-NTA resin for 60 min at 4°C and then packed into a column. The unbound protein was washed with 10 mM Tris with 20 mM imidazole pH 8.0. The bound protein was eluted with increasing concentrations of imidazole (0.01–0.25 M) in 0.02 M PBS (pH 7.5) containing 0.5 M NaCl. The eluted protein was dialyzed against 50 mM Tris, 200 mM KCl, and 5% glycerol (pH 7.5) at 4°C and stored at –70°C.
Site-specific biotinylation of CR2 SR1-4-BCCP using the Bir A enzyme
Site-specific biotinylation was performed following the method of Sarrias et al. (46). To accomplish this, CR2 SCR1-4-BCCP protein was dialyzed against 40 mM Tris containing 5.5 mM MgCl2 and 100 mM KCL (pH 8.0) overnight at 4°C. Following this, 30 µg of SCR1-4-BCCP protein was incubated with 365 µg of purified Bir A protein, 24 µg of d-biotin, and 20 mM ATP for 1 h at 37°C. Excess biotin was removed by a desalting pD-10 column followed by extensive dialysis against PBS. Protein was stored at –70°C in aliquots until use.
SDS-PAGE and Western blot analysis of recombinant proteins
Protein size and purity was determined using nonreducing SDS-PAGE and Western blot analysis. Purified CR2 SCR1-4 and CR2 SCR1-4-BCCP were electrophoresed on a NuPAGE Novex 10% Bis-Tris gel with MOPS NuPAGE as the running buffer. Proteins were then transferred to a nitrocellulose membrane and incubated sequentially with anti-CR2 mAb HB-5 and peroxidase-conjugated goat anti-mouse IgG secondary Ab (Jackson ImmunoResearch Laboratories). In another blot, the same proteins were transferred to a nitrocellulose membrane to evaluate for biotinylation efficiency. In both experiments, nonspecific binding was blocked with PBS 0.05% Tween 20 containing 10% milk. The incorporation of biotin was assessed by reactivity with peroxidase-conjugated SA (BD Pharmingen) at 1 µg/ml. Proteins were detected using the ECL kit (Amersham Biosciences). In addition, N-terminal amino acid sequence analysis was performed for the CR2 SCR1-4 and CR2 SCR1-4-BCCP proteins to confirm the correct protein identity. To accomplish this, a 12% denaturing minigel was prepared and run with SDS-PAGE running buffer containing 100 mM reduced glutathione at 37°C and stored overnight at 4°C before loading the sample. Next, CR2 SCR1-4 and CR2 SCR1-4-BCCP proteins were electrophoresed in running buffer prepared by substituting glutathione with 2-ME. Bands were transferred to polyvinylidene difluoride membrane, and the membrane was stained with 0.1% Coomassie blue in 50% methanol for 5 min. Following slight destaining, the transferred protein bands were excised with a razor blade, and N-terminal sequence analysis was then performed in the Protein Sequencing Core (National Jewish Medical and Research Center, Denver, CO).
Characterization by ELISA of pH and NaCl dependence of CR2 SCR1-15 binding to IFN-
CR2 SCR1-15 was dialyzed overnight at 4°C against 50 mM PBS with differing pH (5.5, 6.0, 6.5, 7.0, 7.5, and 8.0) and in a separate experiment 50 mM PBS containing differing NaCl concentrations (25, 50, 75, 100, 125, and 150 mM). Then, 5 µg/ml human rhIFN- (5 x 106 U/ml) (PBL Biomedical Laboratories) was immobilized on 96-well ELISA plates (Costar) in bicarbonate buffer (pH 9.0) overnight at 4°C. Following washing of the plate, soluble CR2 SCR1-15 dialyzed against differing pH and NaCl buffers was added and incubated for 60 min at 37°C. After washing, CR2 SCR1-15 bound to IFN- was detected using anti-CR2 mAb HB-5, which is specific for SCR3-4, followed by anti-goat mouse IgG (Jackson ImmunoResearch Laboratories) at 1/1000. Color was developed using ABTS, and the OD was measured at an absorbance of 405 nm.
Measurement by ELISA of the interaction of CR2 with the ligands human IFN-, CD23, C3d, and gp350
Dose-dependent binding of CR2 to its ligands was studied using the ELISA method in which 5 µg/ml rhIFN-, CD23, C3d, and gp350 were immobilized on 96-well ELISA plates (Costar) in bicarbonate buffer (pH 9.0) overnight at 4°C. Plates were then washed and blocked with PBS Tween 20 containing 1% BSA. Serial dilutions of CR2 SCR1-15 were then added to the wells in PBS containing 50 mM NaCl (pH 7.4) and allowed to incubate for 1 h binding at 37°C. Following washing, the amount of CR2 SCR1-15 bound to the wells was measured using anti-CR2 mAb HB-5 as described above.
Measurement by surface plasmon resonance of the interaction of CR2 with IFN-, CD23, C3d, and gp350
CR2 binding to its ligands human IFN-, CD23, C3d, and gp350 was studied using BIAcore 3000 (BIAcore) with SA-coupled and carboxymethyl-dextran (CM5) chips. All experiments were performed in 10 mM HEPES, 50 mM NaCl, 1 mg/ml carboxymethyl-dextran, and 0.05% surfactant P-20 (pH 7.4) at 25°C. The chip regeneration buffer used 1 M NaCl in place of 50 mM NaCl. For kinetic analysis, biotinylated CR2 SCR1-4-BCCP (1268 RU) or C3dg-biotin (1221 RU) were immobilized on a SA chip. IFN- (789 RU), CD23 (1228 RU), and gp350 (1466 RU) were immobilized on a CM5 chip using amine-coupling chemistry per manufacturer’s instructions. Immobilization was performed on either flow cell (FC)2 or FC4, and FC1 and FC3 were used as control cells. Binding was measured at 30 µl/min to avoid mass transfer effects. At this flow the initial on-rate was maximum. Flow was allowed for several seconds to establish baseline, and then various concentrations of analyte were injected. The association was allowed for 120 s, and the dissociation of the complex was monitored for 120 s. For IFN- instead of 120 s, 150 s was allowed for both association and dissociation phase to allow optimal binding. BIAevaluation 3.1 (BIAcore) software was used to analyze the binding data, using global fitting where best fit was indicated by a low residual and 
(2) values <10. This analysis yields a Ka, Kd, and KD for protein:protein interactions.
Inhibition of CR2-ligand interaction by mAbs recognizing the first two SCRs of CR2
To determine whether the first two SCR domains were directly involved in CR2 binding to IFN-, CR2 SCR1-15 was preincubated for 1 h at 37°C with IgG1 anti-CR2 mAbs (171, 1048, 994, 629) directed to the SCR1-2 domain (53) or a control IgG1 anti-human factor B mAb (a gift from Dr. J. Thurman, Division of Nephrology and Hypertension, University of Colorado Health Sciences Center, Denver CO). Preincubated CR2 SCR1-15 samples were then added for 1 h at 37°C to an ELISA plate that had been precoated as described above with IFN-, CD23, C3d, and gp350. The amount of CR2 SCR1-15 bound to the wells was detected using the anti-CR2 mAb HB-5 followed by goat anti-mouse IgG (Jackson ImmunoResearch Laboratories) at a 1/1000 dilution. Color was developed using ABTS, and OD measured at an absorbance of 405 nm. Percentage of inhibition was calculated by the following: 1 – [(average experimental OD – background OD)/(maximum average OD – background OD)] x 100.
Cross-competition analysis using CD23, C3d, and gp350
To determine the relationships between the IFN- binding site and other CR2-ligand interaction sites, 2 µg/ml rhIFN- (IFN-; BioSource International) was coated on 96-well ELISA plates (Costar) in bicarbonate buffer (pH 9.0) overnight at 4°C. Plates were then washed and blocked with PBS Tween 20 containing 1% BSA. Serial dilutions of CD23, C3d, and gp350 were made on a diluter plate. CR2 SCR1-15 at 2 µg/ml was then added to the serially diluted wells, as well as to control wells with buffer alone, and incubated for 1 h at room temperature. These samples were then added to IFN--bound ELISA plates and incubated for 1 h. Following washing, the amount of CR2 SCR1-15 bound to the wells was measured using anti-CR2 mAb HB-5 as described above.
B cell culture and stimulation
Human B cells were isolated from healthy donor blood using the RosetteSep Human B cell Enrichment Cocktail (StemCell Technologies). Purified B cells were cultured at a density of 2 x 105/0.1 ml in 96-well flat-bottom plates in RPMI 1640 supplemented with L-glutamine (2 mM), HEPES (20 mM), penicillin (100 U/ml), streptomycin (100 µg/ml), and 10% FBS at 37°C, 5% CO2. B cells were cultured with medium, rhIFN- (IFN-; BioSource International) or IFN- plus anti-CR2 mAb 171 or its isotype control mouse IgG2A (R&D Systems). The B cells were preincubated with Abs for 1 h before adding IFN-. After 6 h of culture, B cells were lysed and stored at –70°C.
RNA extraction and quantitative real-time RT-PCR
RNA was extracted from each lysate using the RNeasy Mini Kit (Qiagen). A total of 0.4 µg of this RNA was reverse-transcribed to cDNA in a 20-µl reaction using SuperScript III RNase H- Reverse Transcriptase (Invitrogen Life Technologies). cDNA obtained from each sample was diluted 1/40, and 10 µl was amplified in a 25-µl real-time PCR using 0.4 µM sense and antisense primers and the 2x iQ SYBR Green Supermix (Bio-Rad). Hypoxanthine guanine phosphoribosyltransferase (HPRT)1 was used as a housekeeping gene. Primer sequences for IFN--inducible genes (IFIG) and HPRT1 were as follows: myxovirus (influenza virus) resistance 1 (MX1) forward, 5'-TACCAGGACTACGAGATTG-3'; MX1 reverse, 5'-TGCCAGGAAGGTCTATTAG-3'; dsRNA-dependant protein kinase (PRKR) forward, 5'-CTTCCATCTGACTCAGGTTT-3'; PRKR reverse, 5'-TGCTTCTGACGGTATGTATTA-3'; HPRT1 forward, 5'- TTGGTCAGGCAGTATAATCC-3'; HPRT1 reverse, 5'- GGGCATATCCTACAACAAAC-3'.
B cells cultured with medium alone were included in each assay to provide a basis for normalization across experiments. Results for each culture condition are expressed as relative expression (RE) compared with B cells cultured with medium. Details of the real-time PCR method have been described in detail (54). Briefly, gene amplifications for each sample were performed in triplicate using the iCycler Real-Time Detection Systems (Bio-Rad). Standard curves for MX1 and PRKR genes were generated using cDNA to determine efficiency. Melting curve analysis was performed for all PCR products to assure specific amplification. To calculate RE in a B cell lysate sample of MX1 and PRKR, each sample was amplified with primers for both the target gene and housekeeping gene in separate wells. Values for target genes MX1 and PRKR and the housekeeping gene HPRT1 were subtracted from the corresponding reference sample. The differences were then used as exponents with the base equal to 1 plus the value of the efficiency of that PCR. The MX1 and PRKR values were divided by the HPRT1 value for each sample, and the result was the relative gene expression of each unknown sample. Results are expressed as percentage of inhibition compared with expression induced by rhIFN- .
Protein determination
The absorption coefficient for each protein calculated at A280 was used to determine protein concentrations.
Statistical analyses
Analyses were performed using Student’s t test.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
Binding sites for the ligands C3d, gp350, and CD23 exist within the SCR1-2 domain of CR2. To study these CR2-ligand interactions in comparison to IFN-, we used two previously described forms of soluble recombinant CR2 containing SCRs 1-2 and 1-15 (Fig. 1), each of which interacts with C3d and gp350 at high affinity (48). In addition, for this current study, we have created a new form of soluble recombinant CR2 containing SCR1-4 (Fig. 1), both with (CR2 SCR1-4-BCCP) and without (CR2 SCR1-4) a biotinylation site at the C terminus. We reasoned that the addition of the SCR3-4 domain would allow us to detect binding of this protein to ligands with the nonligand-blocking mAb HB-5, which reacts with the SCR3-4 domain. In addition, the incorporation of a biotinylation site at the C terminus would allow us to reproducibly orient the ligand binding domain of CR2 to a SA chip, as described previously (46).
We used the mammalian cells NSO and HEK 293 to express CR2 SCR1-4 and CR2 SCR1-4-BCCP, respectively, and proteins were purified using mAb HB-5-Sepharose affinity column. Fig. 2 demonstrates these proteins. Although there are two bands present, 34/35 Kd for CR2 SCR1-4 and 40/41 Kd for CR2 SCR1-4-BCCP, we believe that this minor difference is due to differential glycosylation, because there are two glycosylation sites predicted in this domain, and the differences in apparent m.w. are not present following treatment of the proteins with PN-glycanase (data not shown). In addition, aminoterminal protein sequence analysis demonstrated the identical and expected sequence for each of the four forms (data not shown).
|
The ligands C3d and C3dg-biotin were expressed in E. coli and purified as described previously (6, 48). EBV-gp350 was also expressed and purified using an established method (50). Expression and purification of the lectin-like domain of CD23 is detailed in Materials and Methods.
Identification of optimal pH and NaCl concentration for ligand-binding experiments
We first sought to determine whether CR2 would bind to IFN- by ELISA and to then characterize the optimal binding conditions for further studies. This strategy follows prior experiments in which ELISA has been used to characterize the binding of CR2 to C3d and gp350, and also the finding that optimal binding to C3d occurs at neutral pH and a 50 mM NaCl concentration (46, 48). Fig. 3 demonstrates the results of these analyses. First, CR2 SCR1-15 does clearly bind to IFN- by ELISA (see also Fig. 4A for BSA specificity control). In addition, optimal binding conditions appeared to be very similar to those for C3d, with a decrease in apparent affinity below a pH of 6.5 and above 50 mM NaCl. Because of that, and previous studies in which comparisons were done under neutral pH and low salt conditions (46, 48), all additional experiments were performed in PBS (pH 7.4) and 50 mM NaCl.
|
|
, CD23, C3d, and gp350
Fig. 4 demonstrates the results of experiments in which each of the four ligands, or BSA as a control, is first bound to an ELISA plate. Subsequently, decreasing doses of CR2 SCR1-15 are added, followed by washing and detection of bound CR2 by anti-CR2 mAb HB-5. These results demonstrate a remarkable consistency in which highly specific interactions and dose-dependent interactions are readily detected with each of the four ligands, including IFN-.
Kinetic analysis of the interaction of CR2 SCR1-4 with human IFN-, CD23, C3d, and gp350 by surface plasmon resonance
In this study, we have examined the interaction of CR2 with human IFN-, CD23, C3d, and gp350 by surface plasmon resonance technology using a BIAcore 3000 to compare the binding kinetics of human IFN-, C3d, gp350, and CD23 for surface-attached CR2 SCR1-4. In these experiments, C-terminal biotin-tagged CR2 SCR1-4-BCCP on a SA chip was studied with the ligands human IFN-, CD23, C3d, and gp350 as solution phase analytes. For each of the ligands, the Ka, Kd, and KD were measured and calculated (Fig. 5, A–D, and Table I).
|
|
as a solution phase analyte to the immobilized CR2 SCR1-4-BCCP on a SA chip. We found that the binding reaction was dose-dependent but not saturable at the highest tested concentration of 2000 nM. Similarly, using the same chip, we studied CD23, C3d, and gp350 binding. We found that CD23 binding was dose dependent but, similarly to IFN-, not saturable at the highest tested concentration of 4000 nM. C3d and gp350 demonstrated dose-dependent binding that was saturable at the highest tested concentration of 125 nM. All ligands demonstrated a comparable affinity for CR2 SCR1-4-BCCP, between 263 nM and 467 nM.
We used BIAevaluation software version 3.1 to evaluate whether these ligand-receptor pairs exhibited simple 1:1 binding or whether more than one binding site could be detected. The data analysis revealed that the binding reaction of the experimental curve (gray color) does not fit well to a simple 1:1 Langmuir binding model but rather fits to a two-site binding model (solid lines). These analyses suggest the presence of one high-affinity binding site and a second much lower (2–3 logs or greater difference) affinity binding site in the CR2 SCR1-4 domain. For this interaction, we report only the higher affinity binding site KD values (Table I). 2 values are the goodness of fit parameter describing how precisely the experimental curve fits to the proposed binding model. The 2 values range from 0.46 to 3.0 (Table I) and show that a two-site binding model fits well to the experimental data.
To determine what domain of CR2 is capable of binding to IFN- as well as confirm that the interaction occurs in the reverse orientation during surface plasmon resonance analysis, we used soluble CR2 SCR1-2, SCR1-4, and SCR1-15 with chip-bound IFN-, again in direct comparison to other ligands. Fig. 6 and Table II demonstrate the results of these studies using each of the four chip-bound ligands. When the CR2 SCR1-2 domain that contains known binding sites for the three well-characterized ligands was injected as the solution phase analyte, dose-dependent and saturable binding was observed. Similar dose-dependent binding interactions were seen with both CR2 SCR1-4 and CR2 SCR1-15 for each of the immobilized ligands, including IFN-. However, some differences are present in the Ka and Kd when the ligand is bound by amine coupling chemistry to the CM5 chip as compared with when CR2 SCR1-4-BCCP-biotin is bound to the SA-coated chip. Notably, in this analysis C3d, gp350, and IFN- were found to bind to the solution phase analyte CR2 SCR1-2 at higher affinities when compared with CR2 SCR1-4 and CR2 SCR1-15. In general, in the case of CR2 SCR1-4 and CR2 SCR1-15, the KD decreases as compared with SCR1-2. In contrast, however, CD23 demonstrates a comparatively lower affinity when binding with CR2 SCR1-2 and CR2 SCR1-4 as compared with CR2 SCR1-15. This is likely due to the presence of an additional carbohydrate-depending binding site in CR2 SCR5-8 (4).
|
|
2 values for this proposed binding model fit well, with a range of 0.33–4.0, 0.15–10.0, and 0.25–3.5 for CR2 SCR1-2, SCR1-4, and SCR1-15, respectively. For this interaction (Table II), we report only the higher affinity binding site KD values.
Mapping of IFN- ligand interactions using informative anti-CR2 mAbs
Previous studies using inhibitory mAbs as well as cross-competition of ligands have demonstrated that the binding sites for C3d, gp350, and CD23 overlap within the SCR1-2 domain (4, 53, 55). In the next set of experiments, we determined whether a set of anti-CR2 mAbs generated to and reactive with the SCR1-2 domain that we had previously created (53) would inhibit the binding of IFN- to CR2. As shown in Fig. 7, a very similar pattern of inhibition is found when comparing the relative ability of mAbs to block IFN- binding to CR2 as found when comparing them to C3d (mAb 171>1048>994>629). Importantly, the relative level of inhibition for these four mAbs is identical with that we have previously reported for C3d and gp350 (53). We now show that the other two ligands, CD23 and IFN-, demonstrate a similar pattern of relative inhibition. These results suggest that each ligand binds to a closely related site on CR2 SCR1-2.
|
Previous studies have shown that IFN- may serve to directly block EBV infection of B cells through receptor competition. Our results have suggested that IFN- may as well directly block the binding of other ligands to CR2. To determine whether IFN- directly blocks CR2-ligand interaction, cross-competition binding analyses were performed using IFN--coated wells and determining the effects on the binding of CR2 by adding increasing amounts of CD23, C3d, and gp350 in solution. Fig. 7E demonstrates the results of this analysis and reveals a rank order of competition of gp350>C3d>CD23. This likely reflects the relationships between ligand binding sites, although we cannot rule out a contribution to the results by the relative differences in affinities. Nevertheless, the results are consistent with the presence of closely related binding sites for each ligand.
Identification of IFIG partially inhibited by anti-CR2 mAb 171
Previous studies have reported inhibition of EBV binding to B cells by IFN-, and we have demonstrated in vitro binding in studies shown above. To determine whether there were biologic consequences of the interaction with CR2 on IFIG expression, we used purified B cells and determined using real-time PCR analysis the effects of anti-CR2 mAb 171 on expression of IFIG MX1 and PRKR. We found that both were partially inhibited in a dose-dependent manner by anti-CR2 mAb in comparison to control IgG2a mAb (Fig. 8). Thus, a component of the IFIG response of B cells to IFN- is due to the interaction of this cytokine with CR2.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
binds to CR2 with an affinity that is comparable to three other well-characterized ligands for this receptor studied in parallel and that B cells respond to this by increases in IFIG expression. Each of the four ligands studied interacts with CR2 when either in the fluid phase or when chip-bound. Based on the ability of SCR1-2 to bind IFN- in surface plasmon resonance analysis as well as inhibition experiments with informative mAbs using the full-length CR2 SCR1-15 molecule, it appears that IFN- interacts with CR2 within the same SCR1-2 domain as the other three ligands. As part of these studies, we have also created a novel form of CR2, CR2 SCR1-4-BCCP, that can be readily biotinylated, linked to SA-coated chips, and used in surface plasmon resonance studies in an orientation that mimics the in vivo situation.
We found that human IFN- binds in a dose-dependent manner to surface-attached CR2 SCR1-4-BCCP as well as to CR2 SCR1-2, SCR1-4, and SCR1-15 in solution when IFN- is chip-bound. A slow Ka as well as Kd was observed as compared with C3d and gp350, and the association and dissociation characteristics were observed to be more similar to CD23. Identically to C3d, gp350, and CD23, analysis of the binding data did not fit well to a simple 1:1 Langmuir binding model but rather are more consistent with a two-site binding interaction. However, the high-affinity site is several logs greater than the lower and thus provides the greatest impact on the receptor-ligand interaction.
One interesting but unanticipated aspect of the results is the relative decrease in the apparent affinity of CR2 for IFN- as well as C3d and gp350 as the length of the molecule is increased from SCR1-2 to SCR1-4 and then SCR1-15. This is in contrast to the effects on the apparent affinity for CD23 as the length is increased. We believe this result is due to the presence of a flexible proximal SCR domain structure in solution, a situation that allows the two dispersed binding sites on CD23 in SCR1-2 and SCR5-8 (4) to come together, but one that also partially obstructs SCR1-2 interactions with the ligands that do not use these proximal SCRs (data not shown). This is the subject of ongoing analyses. An alternate explanation we have considered is that the effect is due to differences in glycosylation of the CR2 proteins, because SCR1-2 is deglycosylated before use and CR2 SCR1-4 and CR2 SCR1-15 are not. However, we are unaware of any known effects of differences in glycosylation on direct binding of ligands to the CR2 SCR1-2 domain in solution, because the binding of C3d and gp350 to SCR1-2 (48), as well as CD23 to this domain (4), have been directly shown to be independent of CR2 SCR1-2 glycosylation.
Previous studies using inhibitory mAbs as well as cross-competition of ligands have demonstrated that the binding sites for C3d, gp350, and CD23 overlap within the SCR1-2 domain (4, 53, 55). In this study, we have evaluated whether a set of anti-CR2 mAbs generated to and reactive with the SCR1-2 domain that we had previously created (53) would inhibit the binding of IFN- to CR2. As shown in Fig. 7, a very similar pattern of inhibition is found when comparing the relative ability of mAbs to block IFN- binding to CR2 as found when comparing them to C3d (mAb 171>1048>994>629). There is also direct inhibition of the binding of IFN- to CR2 shown by the other ligands. These results, in conjunction with the observation that CR2 SCR1-2 alone binds IFN- in surface plasmon resonance studies, are most consistent with the binding site for IFN- lying within SCR1-2. However, additional sites of interaction are not ruled out by our studies. Notably, Delcayre et al. (43) reported that mAb HB-5, which interacts with SCRs 3-4, but not SCR1-2-reactive mAb OKB7 blocked IFN- binding to B cells. Therefore, we cannot rule out an important contribution to IFN- binding by regions of CR2 outside of SCR1-2.
IFN- is a member of the type I family of IFNs that are secreted by fibroblasts and lymphocytes, as well as in large quantities by immature plasmacytoid dendritic cells (reviewed in Ref. 56). The type I family consists of structurally homologous cytokines that are involved in the control of viral pathogenesis and cancer, and exhibit both antiviral and antiproliferative effects on cells. Currently, IFN- is also used for treatment of several diseases such as hepatitis C and various types of lymphoid cancers (57).
One well-described complication of the therapeutic administration of IFN- is the development of autoimmunity, especially lupus-like autoantibodies (58). SLE is a chronic autoimmune disease characterized by autoantibodies, the formation of immune complexes, and target organ injury, as well as increased levels of IFN- (59) and an IFN- signature in peripheral blood mRNA (41, 54). Plasmacytoid dendritic cells through the secretion of IFN- can promote the differentiation of B cells into plasma cells in an IL-6-dependent manner (60). In addition, IFN- itself can lead to the activation of B cells in vitro in a mixed PBMC population (61). IFN- can also both promote the development of B cells in the bone marrow as well as enhance BCR-dependent activation events such as calcium flux, proliferation, and induction of activation markers (62).
Inappropriate expression of IFN- from plasmacytoid dendritic cells may play a key role in breaking tolerance in SLE (56, 58). In contrast, whereas the majority of studies suggest that IFN- promotes the development of lupus-like autoimmunity, some caution is necessary because MRL/lpr mice are protected from the development of autoimmunity and target organ injury by type I IFNs (63).
The structure of IFN- family members and their receptor components has been determined by x-ray crystallography and NMR spectroscopy (reviewed in Ref. 64). For example, human IFN-2b consists of an helical structure with a high degree of similarity to other type I IFNs (65). Type I IFNs including IFN- shared a common receptor that consists of two subunits, IFNAR1 and IFNAR2, which associate with each other following ligand binding (66). IFNAR2 exhibits high-affinity binding of type I IFNs with a KD of 
3 nM, which increases 20-fold following association of IFNAR2 into a ternary complex with ligand, whereas the affinity of IFNAR1 alone is lower (>100 nM KD) (66).
A reasonable question that is raised by our studies is what is the biologic role of this IFN- interaction with CR2. There are several possibilities. First, based on our results demonstrating an effect of anti-CR2 mAb on the IFIG response, at least a portion of this response is due to CR2, and it is possible that some genes are induced in B cells by IFN- solely or preferentially through CR2. It is relevant to point out that several IFN--dependent activation events in primary human B cells have not been completely inhibited when cells are treated with an inhibitory anti-IFNAR-1 mAb (61), which is consistent with our results. Second, IFN- binding may impart a BCR-independent signal through CR2 in a similar fashion as previously described for other ligands, where homotypic adhesion (18), NF-B activation (19), IL-6 generation (20), and actin polymerization (23) have been described as potential outcomes. Third, IFN- may block the effects of CD23 as it binds to CR2 on B cells and promotes isotype switching to IgE (3). Fourth, as previously suggested, IFN- may serve to directly block EBV infection of B cells through receptor competition (43), or may block the binding of other ligands to CR2. Consistent with this possibility, experiments using ELISA have shown that IFN- can block other ligands from interacting with CR2. And finally, although it is unlikely that IFN- would coligate CR2 with the BCR similarly to C3d-bound Ags, down-modulation of BCR-dependent signals through sequestration of the CR2-associated molecule CD19 (67) may occur and interfere with CD19-dependent BCR signal amplification (68). Each of these possibilities is a testable hypothesis for future experiments.
An additional important question for further analysis is whether IFN- binds to mouse CR2. This is relevant because, whereas C3-derived ligands bind mouse and human CR2 comparably, gp350 does not bind mouse CR2 (54, 69), and mouse CD23 has never been convincingly shown to bind mouse CR2. Thus, whether IFN- is similar to the former or latter situation is important to resolve.
Finally, these data again point to intriguing relationships between CR2 and SLE. As noted above, decreased levels of CR2 appear to promote loss of tolerance in this disease. In addition, prior infection with EBV, another CR2 ligand, has been closely linked with the development of SLE (70), and increased serum levels of EBV have been identified in patients with SLE (71). We now confirm in this study that IFN-, also associated with SLE, is another ligand for CR2 and that a portion of the IFIG response that is linked to SLE is due to CR2 interactions. In toto, these observations continue to support the hypothesis that CR2 is a key contributor to the pathogenesis of SLE, potentially through several molecular and immune mechanisms.
|
Acknowledgments |
|---|
|
Disclosures |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
|
Footnotes |
|---|
1 This work was supported by the Lupus Research Institute (to V.M.H., M.K.C., and R.A.), National Institutes of Health Grants R0-1 CA53615 (to V.M.H.) and R0-1 AR050829 (to M.K.C.), and the Alliance for Lupus Research (to M.K.C.). 
2 Address correspondence and reprint requests to Dr. V. Michael Holers, Division of Rheumatology, Department of Medicine, Box B-115, University of Colorado Health Sciences Center, 4200 East Ninth Avenue, Denver, CO 80262. E-mail address: Michael.Holers{at}UCHSC.edu
3 Abbreviations used in this paper: CR, complement receptor; SCR, short consensus repeat; SLE, systemic lupus erythematosus; IFNAR, IFN- receptor; BCCP, biotin carboxyl carrier protein; BSP, biotinylation signal peptide; IPTG, isopropyl -D-thiogalactoside; SA, streptavidin; rh, recombinant human; FC, flow cell; HPRT, hypoxanthine guanine phosphoribosyltransferase; IFIG, IFN--inducible gene; MX1, myxovirus (influenza virus) resistance 1; PRKR, dsRNA-dependant protein kinase; RE, relative expression.
Received for publication October 24, 2005. Accepted for publication April 13, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionDisclosuresReferences |
|---|
B and different signaling pathways. J. Mol. Biol. 308: 501-514. [Medline]
