C3d can function as a molecular adjuvant by binding CD21 and thereby enhancing B cell activation and humoral immune responses. However, recent studies suggest both positive and negative roles for C3d and the CD19/CD21 signaling complex in regulating humoral immunity. To address whether signaling through the CD19/CD21 complex can negatively regulate B cell function when engaged by physiological ligands, diphtheria toxin (DT)-C3d fusion protein and C3dg-streptavidin (SA) complexes were used to assess the role of CD21 during BCR-induced activation and in vivo immune responses. Immunization of mice with DT-C3d3 significantly reduced DT-specific Ab responses independently of CD21 expression or signaling. By contrast, SA-C3dg tetramers dramatically enhanced anti-SA responses when used at low doses, whereas 10-fold higher doses did not augment immune responses, except in CD21/35-deficient mice. Likewise, SA-C3dg (1 μg/ml) dramatically enhanced BCR-induced intracellular calcium concentration ([Ca2+]i) responses in vitro, but had no effect or inhibited [Ca2+]i responses when used at 10- to 50-fold higher concentrations. SA-C3dg enhancement of BCR-induced [Ca2+]i responses required CD21 and CD19 expression and resulted in significantly enhanced CD19 and Lyn phosphorylation, with enhanced Lyn/CD19 associations. BCR-induced CD22 phosphorylation and Src homology 2 domain-containing protein tyrosine phosphatase-1/CD22 associations were also reduced, suggesting abrogation of negative regulatory signaling. By contrast, CD19/CD21 ligation using higher concentrations of SA-C3dg significantly inhibited BCR-induced [Ca2+]i responses and inhibited CD19, Lyn, CD22, and Syk phosphorylation. Therefore, C3d may enhance or inhibit Ag-specific humoral immune responses through both CD21-dependent and -independent mechanisms depending on the concentration and nature of the Ag-C3d complexes.
A role for complement C3 cleavage products as regulators of B cell signaling and humoral immunity has been appreciated for decades (1, 2). Activated C3 forms covalent bonds with foreign Ags and generates C3dg and C3d cleavage products that can supply physiologically relevant ligands for the human CD21 and mouse CD21/35 complement receptors expressed by mature B cells and follicular dendritic cells (3, 4). C3dg and C3d also function as natural adjuvants to enhance humoral immune responses to Ags, including pneumococcal capsular polysaccharide, hen egg lysozyme (HEL),5 streptavidin (SA), HIV gp120, and influenza or measles virus hemagglutinin (5, 6, 7, 8, 9, 10). Moreover, deficiencies in either C3 or the common gene that generates C3 complement receptors 1 (CD35) and 2 (CD21) result in impaired Ab responses in mice (11, 12, 13, 14, 15, 16, 17). However, complement components may also exhibit immunosuppressive activities (18). For example, C3 depletion by cobra venom factor can inhibit primary, yet enhance secondary, humoral immune responses to both unmodified and protein-conjugated pneumococcal polysaccharide (19). Furthermore, C3d attachment to certain Ags can inhibit humoral responses to the nontoxic diphtheria toxin fragment B (DT), human chorionic gonadotropin, bovine rotavirus VP7, bovine herpesvirus type 1 glycoprotein D, and malarial circumsporozoite protein (20, 21, 22, 23). Thus, in addition to its positive role in regulating immune responses, complement may negatively regulate humoral immunity through unknown mechanisms.
CD21 is a component of a multimeric complex on B cells (24, 25) that includes CD19, a molecule that generates transmembrane signals during B cell activation (26). Co-cross-linking of the CD19/CD21 complex with the BCR complex using either C3d-tagged Ags or CD19 mAbs can augment BCR-induced [Ca2+]i responses and lower B cell activation thresholds (5, 27, 28, 29). SA-C3dg tetramers, formed by SA binding four monobiotinylated C3dg molecules, also augments B cell [Ca2+]i responses and specifically binds CD21 on wild-type, but not CD21−/−, B cells (13, 30). Anti-IgM and SA-C3dg complexes formed by SA binding of biotinylated anti-IgM Abs and monobiotinylated C3dg molecules also augment B cell [Ca2+]i responses (30, 31). Consistent with CD21 signaling through CD19, B cells from CD19-deficient (CD19−/−) mice are hyporesponsive to most transmembrane signals (32, 33, 34), whereas B cells from transgenic mice that overexpress CD19 are hyper-responsive (34, 35, 36). These findings support the concept that CD19/CD21 engagement by C3d positively regulates BCR signal transduction.
Although C3 cleavage products can positively regulate B cell function and humoral immunity through the CD19/CD21 complex, C3dg can also function as a molecular adjuvant for SA through CD21-independent pathways (10). In addition, CD21 engagement can negatively regulate B cell function. Excess CD21 ligation using mAbs can inhibit B cell [Ca2+]i responses (37) and proliferation (38). Monovalent CD21 ligands can also inhibit anti-Ig-induced [Ca2+]i responses in human B cells (39). Likewise, mAb ligation of CD19 can suppress B cell [Ca2+]i responses and/or proliferation after BCR ligation (40, 41, 42, 43). More recently, mAb ligation of CD19 was found to either augment or inhibit BCR-induced [Ca2+]i responses depending on the concentration of CD19 mAb used (44). CD19 augmentation or inhibition of BCR-induced [Ca2+]i responses does not require CD19 and BCR complex co-cross-linking, but does require their simultaneous ligation (44). Thus, it is possible that CD19/CD21 ligation can exert negative regulatory roles, with Ag/complement complexes down-regulating B cell activation to limit humoral immune responses. This hypothesis is attractive given that a loss of CD21 function correlates with autoimmune disease progression in mice (45, 46, 47). However, because all previous studies of B cell negative regulation through the CD19/CD21 complex have relied on mAbs for CD19/CD21 complex engagement, it remains unknown whether CD21 ligation by physiological ligands can down-regulate B cell activation or limit humoral immune responses. Therefore, this study examined whether ligation of the CD19/CD21 complex using two structurally distinct model Ag/complement complexes that bind CD21 could enhance or inhibit humoral immune responses and BCR-induced [Ca2+]i responses through CD21-dependent processes. The results demonstrated that Ag-complement complexes can be both stimulatory and inhibitory depending on the extent of CD21 ligation by either augmenting or inhibiting multiple BCR-induced signaling pathways.
Materials and Methods
CD21/35−/− (13), CD19−/− (32), and Lyn−/− (48) mice were backcrossed with C57BL/6 wild-type mice (The Jackson Laboratory) for seven or more generations. All mice were housed in a specific pathogen-free barrier facility and were used for experiments at 2–3 mo of age. All procedures were approved by the Duke University Animal Care and Use Committee.
Recombinant biotinylated C3dg was produced and purified as previously described (30). SA-C3dg tetramers were generated by mixing biotinylated C3dg (80 μg) with SA (20 μg; Sigma-Aldrich) in 200 μl of PBS for 30 min at room temperature before appropriate dilution and use for experiments within 2 h. Plasmids encoding the nontoxic B fragment of DT and DT fused to three tandem repeats of C3d (DT-C3d3) were generated using aa 1006–1303 of C3 and were purified and used as previously described (22).
Mouse immunizations and ELISAs
For SA immunizations, mice were injected i.p. with SA (10 or 100 μg) or SA complexed with biotinylated C3dg in 200–500 μl of PBS. C3dg was treated with polymyxin B-agarose (Sigma-Aldrich) before SA-C3dg tetramer formation. Tetramers were generated as described above, but were formed in half the volume and subsequently diluted in PBS immediately before immunizations. For DT immunizations, mice were injected with 250 pmol of recombinant DT or DT-C3d3 fusion protein i.p. in 200 μl of PBS and boosted with the same quantity of Ag on day 14 after the initial injections.
Ag-specific Abs were quantified by coating wells in 96-well plates with SA (5 μg/ml; 100 μl/well), recombinant C3dg (5 μg/ml; 100 μl/well), or recombinant DT or DT-C3d3 (25 pmol/100 μl/well) in 0.1 M borate-buffered saline overnight at 4°C. Plates were washed in TBS and blocked with TBS containing 1% gelatin and 2% BSA for 90 min at 37°C. Sera were diluted 1/100 (SA) or 1/200 (DT) in TBS containing 1% BSA and incubated in duplicate wells at room temperature for 90 min. Plates were washed with TBS containing 1% Tween 20 and incubated with alkaline phosphatase-conjugated polyclonal goat anti-mouse IgM, IgG, or Ig Abs (Southern Biotechnology Associates) for 1 h at room temperature. Plates were developed using p-nitrophenyl phosphate substrate (Southern Biotechnology Associates), and OD405 values were determined.
MB19-1 mouse anti-mouse CD19 mAb (IgA) (36) was affinity purified (KAPTIV-AE; Tecnogen) from salt-fractionated ascites fluid. Other Abs used include the 1D3 rat anti-mouse CD19 (IgG) and FITC- or CyChrome-conjugated anti-B220 (RA3-6B2) mAbs (BD Pharmingen), FITC-conjugated anti-mouse IgA Ab (Sigma-Aldrich), FITC-conjugated anti-SA (Vector Laboratories), goat F(ab′)2 anti-mouse IgM Ab (Cappel Laboratories and ICN Biomedicals), and HRP-conjugated anti-phosphotyrosine (anti- pTyr) Ab (4G10; Upstate Biotechnology). Anti-pTyr (PY99), anti-Lyn (Lyn 44), anti-Syk (LR and C-20), anti-Src homology 2 domain-containing protein tyrosine phosphatase-1 (anti-SHP-1; C-19), and anti-ERK2 (C-14) Abs were purchased from Santa Cruz Biotechnology. Abs specific for phosphorylated proteins, including anti-phospho-Src family kinases (Tyr416), anti-phospho-CD19 (Tyr513), anti-phospho-Akt (Ser473), anti-phospho-phospholipase Cγ1 (anti-phospho-PLCγ1; Tyr783), and anti-phospho-p38 (Tyr180/Tyr182) were obtained from Cell Signaling Technologies. The CD22-specific MB22-1 (49) and MB22-8 (binds extracellular domains) mAbs were used as hybridoma tissue culture supernatants. Rabbit antisera generated against the CD19 cytoplasmic domain (provided by Dr. M. Grove, Duke University, Durham, NC) were previously described (50). Immunofluorescence staining for cell surface CD19 expression with flow cytometric analysis was performed as described previously (51).
SA-C3dg binding to spleen B cells
Splenocytes (107/ml) were incubated with SA-C3dg or equivalent amounts of SA in 100 μl of PBS containing 2% FBS for 30 min on ice. The cells were washed twice in cold PBS plus 2% FBS, incubated with FITC-conjugated anti-SA Ab (Vector Laboratories; 1/250 dilution) for 30 min on ice, washed, and incubated with CyChrome-conjugated anti-B220 Ab (BD Pharmingen; 1/800 dilution) on ice for 30 min. After washing twice in cold PBS plus 2% FBS, the cells were fixed in 400 μl of PBS containing 1.5% formaldehyde and analyzed by flow cytometry on a FACScan flow cytometer (BD Biosciences) using CellQuest software with fluorescence intensities shown on a 4-decade log scale. For fluorescence intensity determinations, 10,000 B220+ gated events were collected. The levels of background staining obtained using SA only and FITC-conjugated anti-SA Ab were subtracted from SA-C3dg staining levels to obtain specific SA binding mediated by C3dg.
Splenocytes (107/ml) were loaded with 1 μM indo-1/AM ester (Molecular Probes) at 37°C for 20 min in RPMI 1640 medium containing 5% (v/v) FBS and 10 mM HEPES as previously described (44). The cells were washed and stained with FITC-conjugated anti-B220 Ab (BD Pharmingen) for 10 min at 37°C. After washing twice with RPMI 1640 medium containing 5% (v/v) FBS and 10 mM HEPES, the cells were resuspended in the same medium (2 × 106/ml) and protected from light at room temperature (<2 h) until flow cytometric analysis. The ratio of fluorescence (405/525 nm) for B220+ cells was measured using a FACStar flow cytometer (BD Pharmingen) with or without stimuli after 1-min incubations for baseline determinations. Results were plotted as fluorescence ratios with the baselines (average of 0–60 s) subtracted. Increased fluorescence ratios indicate increased [Ca2+]i.
B cell activation, immunoprecipitations, and Western blot analysis
B cell-enriched, single-cell splenocyte preparations were generated by incubating 2 × 108 splenocytes with 180 μl of anti-Thy1.2 Ab-coated magnetic beads (Dynal Biotech) in 10 ml of RPMI 1640 medium containing 5% FBS for 30 min at 4°C, followed by T cell removal using a magnet. B cell preparations were ≥93% B220+ as determined by immunofluorescence staining with flow cytometric analysis. B cells suspended (2 × 107/ml) in 500 μl of RPMI 1640 medium containing 5% FBS at 37°C were incubated with goat F(ab′)2 anti-mouse IgM Ab alone or in combination with SA-C3dg or MB19-1 mAb as indicated. The cells were subsequently lysed in 300 μl of buffer containing 1% Nonidet P-40 (Igepal; Sigma-Aldrich), 150 mM NaCl, 50 mM Tris-HCl (pH 8.0), 1 mM sodium orthovanadate, 2 mM EDTA, 50 mM sodium fluoride, and 1 mM PMSF. For immunoprecipitations, cell lysates were incubated with the appropriate Abs for 1 h, followed by a 2-h incubation with protein G-Sepharose beads (30 μl; Amersham Biosciences) at 4°C. For CD19 immunoprecipitations using the MB19-1 or control mAb, the lysates were incubated with 30 μl of MB19-1 mAb conjugated to Affigel 10 beads (1 mg of MB19-1 mAb/ml beads; Bio-Rad). Pelleted beads were washed four times with ice-cold lysis buffer. After separation by SDS-PAGE, the immunoprecipitated proteins were transferred onto nitrocellulose membranes (Schleicher & Schuell) for Western blot analysis. The membranes were incubated with either anti-pTyr-specific Ab (PY99) or Abs specific for the protein of interest (1/1,000), followed by incubation with HRP-conjugated anti-rabbit or anti-mouse IgG secondary Abs (1/20,000; Jackson ImmunoResearch Laboratories). The blots were developed using an ECL kit (Pierce). To verify equal amounts of protein in each lane, the blots were stripped by incubation with 0.625 M Tris buffer containing 2% SDS and 0.7% 2-ME (pH 6.4) for 30 min at 65°C and reprobed with Abs specific for ERK2 or the protein of interest. For quantification of band intensities, autoradiographs were scanned (Scanmaker E6; Microtek) and analyzed using National Institutes of Health Image software (version 1.60).
Results between experiments were compared using autoradiographs with similar, if not identical, exposure times. Band intensities for the experimental results and control proteins within each Western blot were first quantified and expressed as software-defined pixel densities to generate relative unit measurements. Second, small variations in protein loading between lanes were normalized by comparing control protein band densities relative to control band densities at time zero. Third, the band intensities for the experimental results were adjusted to compensate for relative variations in protein loading between lanes. Then, the arbitrary unit values from all experiments were pooled and analyzed.
Student’s t test was used to determine the significance of differences between sample means.
Effects of C3d-Ag complexes on Ag-specific immune responses
Two distinct model Ag/complement complexes were used, DT-C3d3 fusion protein and SA-C3dg tetramers, because DT elicits robust Ab responses, whereas SA induces modest Ab responses. DT-C3d3 fusion protein was generated as previously described (22) using procedures similar to those reported for HEL-C3d3 fusion protein generation (5). When purified, recombinant DT (∼41 kDa) and DT-C3d3 (∼141 kDa) represented the vast majority of protein in each preparation (Fig. 1,A). Wild-type mice were immunized with purified recombinant DT and DT-C3d3 fusion proteins and boosted on day 14; DT-specific IgG responses were examined on day 28. DT-immunized mice generated significant IgG responses that were reactive with both DT and DT-C3d3 fusion proteins, but not with recombinant C3dg (Fig. 1,B). By contrast, DT-C3d3-immunized mice generated modest DT-specific IgG responses that were >9-fold lower than those of DT-immunized littermates. Sera from DT-immunized mice reacted specifically and equally with DT and DT-C3d3 in Western blot assays, whereas sera from DT-C3d3-immunized mice were minimally reactive with DT, DT-C3d3, or C3dg (Fig. 1,C). That anti-DT sera showed similar reactivity with DT and DT-C3d3 indicates that C3d addition did not alter Ab-binding epitopes on DT protein or their accessibility to Ab binding. C3d3 inhibition of DT-specific immune responses was specific, because DT-C3d3 immunizations did not inhibit immune responses to other simultaneously administered Ags (data not shown). DT-C3d3 binding to B cells is mediated by CD21, because the 7G6 Ab that blocks the C3d binding site of CD21 (7G6) significantly reduces DT-C3d3 binding (22). Likewise, we were unable to detect significant DT-C3d3 binding to CD21−/− B cells (data not shown). To determine whether C3d inhibition of DT immune responses was CD21 dependent, wild-type and CD21/35−/− mice were immunized with DT or DT-C3d3. DT immunization induced IgM and IgG responses in both wild-type and CD21/35−/− mice (Fig. 1 D), although DT-specific IgG responses were lower in CD21/35−/− mice by day 28, consistent with CD21 deficiency impairing humoral immune responses (10, 11, 12, 13). However, DT-C3d3 immunization failed to induce significant DT-specific IgM or IgG responses in either wild-type or CD21/35−/− mice. Thus, C3d attachment to DT dramatically inhibited the induction of DT-specific humoral immune responses independently of CD21 expression.
SA-C3dg was generated by combining recombinant biotinylated C3dg (∼45 kDa) with SA (∼52–66 kDa) in an ∼4:1 molar ratio as previously described (10). Because the stoichiometry of SA-C3dg complexes remains theoretical, and the size of SA varies (52), when indicated we cite the concentration of C3dg used to generate protein complexes. Spleen B cells bound SA-C3dg at levels significantly above background when used at concentrations as low as 1 μg/ml (p < 0.05), whereas binding did not reach saturating levels when used at concentrations as high as 200–400 μg/ml (Fig. 2,A). B cell binding to SA-C3dg required CD21 expression (data not shown) as previously described (13, 30). To examine the adjuvant effect of C3dg in vivo, wild-type mice were immunized with either low (10 μg) or high (100 μg) SA protein doses alone or complexed with 40 and 400 μg of C3dg, respectively. Low-dose SA immunizations did not induce significant SA-specific IgM or IgG responses in wild-type mice, whereas the addition of C3dg to low-dose SA induced ≥7-fold higher IgM and IgG responses (Fig. 2 B). In fact, anti-SA Ab responses to low-dose SA-C3dg were similar to those induced by high-dose SA alone. Addition of C3dg to high-dose SA induced 1.9- and 2.5-fold higher IgM and IgG responses, respectively. Thus, C3dg functioned as an adjuvant with low-dose SA in wild-type mice, but did not enhance SA-specific immune responses to the same extent when SA and C3dg were used at 10-fold higher concentrations.
To determine whether low- or high-dose SA-C3dg immune responses were CD21 dependent, CD21/35−/− mice were immunized i.p. with SA or SA-C3dg. Low-dose SA immunizations did not induce significant SA-specific IgM or IgG responses in CD21/35−/− mice by day 14 (Fig. 2 C). However, immunization of CD21/35−/− mice with either low-dose SA-C3dg or high-dose SA induced measurable Ab responses. Surprisingly, high-dose SA-C3dg immunizations in CD21/35−/− mice generated 5- to 9-fold higher Ab responses than either low-dose SA-C3dg or high-dose SA. Thus, C3dg functioned as a molecular adjuvant for low- and high-dose SA in the absence of CD21 expression. These results suggest that the adjuvant activity of C3d is dependent on the nature and dose of the C3d-Ag complexes and may function through both CD21-dependent and -independent pathways. Moreover, the finding that CD21/35−/− mice produced significantly higher Ab responses (≥5-fold higher) to high-dose SA-C3dg than to high-dose SA, whereas wild-type mice immunized with high-dose SA or high-dose SA-C3dg generated similar Ab responses, suggests that CD21 expression may actually inhibit SA-specific Ab responses to high-dose SA-C3dg.
Effects of Ag-C3d complexes on BCR-induced [Ca2+]i responses
To examine whether C3d could influence immune responses by affecting CD19/CD21 complex signaling, the effects of DT-C3d3 and SA-C3dg on BCR-induced [Ca2+]i responses were measured. Goat F(ab′)2 anti-mouse IgM Ab at 40 μg/ml was determined empirically to induce optimal B cell [Ca2+]i responses (data not shown). The simultaneous addition of DT-C3d3 did not significantly affect BCR-induced [Ca2+]i responses in spleen B cells at any concentration assessed (0.1–100 μg/ml; Fig. 3,A and data not shown). DT-C3d3 also failed to significantly affect suboptimal (10 μg/ml) anti-IgM Ab-induced [Ca2+]i responses (data not shown). By contrast, SA-C3dg at 0.2–5 μg/ml augmented BCR-induced [Ca2+]i responses (Fig. 3,B). SA-C3dg at 10 μg/ml did not significantly influence [Ca2+]i responses, whereas CD21 ligation using SA-C3dg at 50 μg/ml or higher concentrations slowed the initial rate of increase and diminished the magnitude of BCR-induced [Ca2+]i responses. The reciprocal stimulatory and inhibitory influences of SA-C3dg were also observed when B cells were activated using suboptimal (10 μg/ml) concentrations of anti-IgM Abs (Fig. 3,C). SA-C3dg did not affect BCR-induced [Ca2+]i responses in CD21−/− or CD19−/− B cells (Fig. 3, D and E), confirming a role for the CD19/CD21 complex in both the stimulatory and inhibitory activities of C3dg. Likewise, CD19/CD21 complex ligation using a mouse IgA CD19 mAb at 10 μg/ml augmented BCR-induced [Ca2+]i responses, but inhibited [Ca2+]i responses when used at 40 μg/ml (Fig. 3,F), as previously described (44). The effects of CD19 mAb on BCR-induced [Ca2+]i responses were characteristically more dramatic than those of SA-C3dg, consistent with CD19 being expressed at higher densities than CD21 on the cell surface (53). CD19 mAb treatment of B cells alone induced characteristic low-level [Ca2+]i responses, whereas DT, DT-C3d3, SA-C3dg, C3dg monomers, SA, control mouse IgA, or buffer alone did not induce detectable [Ca2+]i responses in these assays (Fig. 3 G and data not shown). Thus, CD19/CD21 complex ligation by C3dg qualitatively influenced BCR-induced [Ca2+]i responses in a dose-dependent manner without a requirement for BCR-CD19/CD21 co-cross-linking.
Effects of Ag-C3d complexes on BCR-induced signaling
To determine how SA-C3dg binding influenced BCR-induced [Ca2+]i responses, alterations in CD19 phosphorylation were assessed after simultaneous BCR and CD21 or CD19 ligation. B cell stimulation with anti-IgM Abs (40 μg/ml) induced rapid CD19 phosphorylation that was maximal after 1 min (Fig. 4, A and B). The simultaneous addition of SA-C3dg (1 μg/ml) significantly enhanced the magnitude of BCR-induced CD19 phosphorylation, with maximal CD19 phosphorylation induced after 20–40 s of stimulation (Fig. 4,A). Likewise, CD19 mAb (10 μg/ml) addition significantly enhanced the magnitude of BCR-induced CD19 phosphorylation, with maximal CD19 phosphorylation induced as early as 20 s after stimulation (Fig. 4,B). By contrast, CD21 ligation by high concentrations of SA-C3dg (50 μg/ml) did not alter the kinetics or intensity of BCR-induced CD19 phosphorylation (Fig. 4,A). CD19 ligation by high concentrations of CD19 mAb (40 μg/ml) significantly inhibited and delayed CD19 phosphorylation after BCR ligation (Fig. 4 B). Thus, C3d differentially regulated BCR-induced CD19 phosphorylation depending on the extent of CD19/CD21 complex engagement.
CD21 engagement by SA-C3dg (1 μg/ml) significantly augmented overall cellular protein tyrosine phosphorylation after BCR engagement (Fig. 4,C) and significantly augmented BCR-induced phosphorylation of Lyn and other Src family protein tyrosine kinases, PLCγ, Akt, and p38 kinase (Fig. 4,D). High concentrations of SA-C3dg (50 μg/ml) also enhanced overall BCR-induced protein tyrosine phosphorylation, but at levels lower than those observed with SA-C3dg at 1 μg/ml (Fig. 4,C). Notably, SA-C3dg binding (50 μg/ml) reduced the phosphorylation of ∼150- and ∼70-kDa proteins below the levels induced by anti-IgM alone (Fig. 4, arrowheads). High levels of SA-C3dg did not significantly augment BCR-induced phosphorylation of Lyn and other Src family protein tyrosine kinases, PLCγ, Akt, and p38 kinase (Fig. 4 D). SA-C3dg binding alone did not induce tyrosine phosphorylation of spleen B cell proteins or affect Lyn, PLCγ, Akt, or p38 kinase phosphorylation (data not shown). Thus, CD21 ligation by SA-C3dg at low concentrations significantly augmented BCR-induced cellular protein tyrosine phosphorylation and phosphorylation of important signaling proteins, although higher levels of CD21 ligation had minimal effects.
Role of Lyn in CD19 phosphorylation
We have previously shown that Lyn activity is crucial for CD19 phosphorylation after BCR engagement (51) and that CD19 amplifies Lyn activity through processive amplification (51, 54). By contrast, Xu et al. (55) have reported that CD19 phosphorylation is normal in Lyn-deficient (Lyn−/−) B cells and proposed that CD19 and Lyn function independently. Our studies relied primarily on the MB19-1 mAb (36), whereas Xu et al. (55) used the 1D3 (rat IgG) CD19 mAb (56). To resolve this issue regarding Lyn’s role in CD19 function, CD19 phosphorylation was reassessed in B cells from CD19−/− and Lyn−/− mice using MB19-1 and 1D3 mAbs in side-by-side comparisons. Both MB19-1 and 1D3 mAbs immunoprecipitated phosphorylated CD19 from wild-type B cells after BCR engagement (Fig. 5,A). Although immunoprecipitated CD19 is normally resolved as a broad 90,000–110,000 kDa band after SDS-PAGE analysis due to its extensive glycosylation and complex phosphorylation patterns (Figs. 5 and 6), CD19 was visualized as a narrow distinct band in the studies by Xu et al. (55). The 1D3 mAb also immunoprecipitated a distinct ∼120-kDa tyrosine-phosphorylated protein (Fig. 5, A and C, upper panel, arrowhead) that was not precipitated by the MB19-1 mAb (Fig. 5,A). Furthermore, the 1D3 mAb immunoprecipitated the same 120-kDa tyrosine-phosphorylated protein from CD19−/− B cells. Immunoprecipitations under the same conditions with control Abs did not isolate the 120-kDa protein. Likewise, the 120-kDa phosphoprotein immunoprecipitated by 1D3 mAb did not react with polyclonal CD19 cytoplasmic domain Abs (Fig. 5, A and C, lower panel), suggesting that this phosphoprotein is not CD19. Furthermore, tyrosine phosphorylation of the ∼120-kDa protein varied between experiments relative to the level of CD19 phosphorylation, with the range of variability reflected in the two experiments shown in Fig. 5 A. Thus, the 1D3 mAb reacted with a 120-kDa phosphoprotein in addition to cell surface CD19.
CD19 phosphorylation was also examined in splenic B cells from wild-type and Lyn−/− mice after BCR engagement using a phosphoCD19-specific Ab. Although phosphorylation of CD19, Lyn and other Src family members was induced in wild-type B cells after BCR engagement, Lyn phosphorylation was not observed, and CD19 phosphorylation was modest in Lyn−/− B cells (Fig. 5, B and C). Identical results were obtained with an anti-phosphotyrosine mAb after CD19 immunoprecipitation by MB19-1 mAb (data not shown) (51). By contrast, the ∼120-kDa tyrosine-phosphorylated protein immunoprecipitated by 1D3 mAb was phosphorylated in both wild-type and Lyn−/− B cells (Fig. 5,C). Because spleen B cells from wild-type and Lyn−/− mice expressed cell surface CD19 at identical levels (Fig. 5 D), the absence of CD19 phosphorylation in Lyn−/− B cells most likely results from the absence of Lyn. Thus, Lyn expression was crucial for CD19 phosphorylation, and discrepancies in previous publications may have resulted from the use of CD19 mAbs with differing specificities.
C3dg binding augments Lyn interactions with CD19 after BCR ligation
Because Lyn phosphorylates and binds to phosphoCD19 in vitro and in vivo after BCR ligation (51), the effect of SA-C3dg binding on the association of Lyn with CD19 was assessed. SA-C3dg binding (1 μg/ml) significantly increased the amount of Lyn coimmunoprecipitated with CD19 after BCR ligation (Fig. 6,A). By contrast, SA-C3dg at 50 μg/ml did not enhance Lyn/CD19 associations after BCR ligation. Similarly, Lyn binding to CD19 after BCR ligation was significantly increased by MB19-1 mAb (10 μg/ml) ligation of CD19 (Fig. 6 B). However, there was no augmentation of BCR-induced Lyn/CD19 associations after CD19 ligation using MB19-1 mAb at 40 μg/ml. Thus, increased and accelerated BCR-induced CD19 phosphorylation by optimal CD19/CD21 complex ligation was probably due to increased Lyn recruitment to CD19. By contrast, increased Lyn/CD19 associations were not observed when the CD19/CD21 complex was ligated to a greater extent.
C3dg regulation of Syk phosphorylation after BCR ligation
The effect of C3dg binding on Syk tyrosine phosphorylation was examined, because Syk phosphorylation is crucial for BCR-induced [Ca2+]i responses (57, 58). SA-C3dg (1 μg/ml) alone had no measurable effect on BCR-induced Syk phosphorylation at either 20 s or 3 min after BCR stimulation (Fig. 6,C). However, CD21 ligation by SA-C3dg at 50 μg/ml significantly reduced BCR-induced Syk phosphorylation at both time points. Similarly, CD19 ligation by MB19-1 mAb at 10 μg/ml had almost no effect on Syk phosphorylation after BCR ligation, whereas CD19 ligation by MB19-1 mAb at 40 μg/ml significantly reduced BCR-induced Syk phosphorylation at both 20 s and 3 min after stimulation (Fig. 6 D). Thus, C3dg ligation of the CD19/CD21 complex may inhibit BCR-induced [Ca2+]i responses by down-regulating Syk phosphorylation.
C3dg regulation of CD22 phosphorylation
CD19 and CD22 are components of a reciprocal regulatory loop that influences BCR-induced [Ca2+]i responses (44, 59, 60). Therefore, the possibility that SA-C3dg binding to the CD19/CD21 complex regulates BCR-induced [Ca2+]i responses by affecting CD22 function was examined. CD22 phosphorylation increased as early as 20 s after BCR ligation, continued to increase until at least 3 min after stimulation, and decreased thereafter (data not shown). CD21 ligation by 1 μg/ml SA-C3dg reduced BCR-induced CD22 phosphorylation slightly by 3 min after stimulation, whereas SA-C3dg binding at 50 μg/ml significantly reduced CD22 phosphorylation by 3 min (Fig. 7,A). Tyrosine-phosphorylated CD22 recruits SHP-1 phosphatase, a negative regulator of B cell activation (61, 62, 63). However, SA-C3dg binding at concentrations of both 1 and 50 μg/ml reduced SHP-1 associations with CD22 by 3 min after BCR ligation (Fig. 7,A). Similarly, CD19 ligation by MB19-1 mAb at both 10 and 40 μg/ml significantly reduced BCR-induced CD22 tyrosine phosphorylation and SHP-1/CD22 associations at 3 min (Fig. 7 B). Thus, SA-C3dg binding negatively influenced CD22 phosphorylation and its recruitment of SHP-1.
The current results indicate that the CD19/CD21 complex functions as a general innate sensing mechanism that can positively and negatively influence B cell signaling and function. Specifically, two model Ag-C3d complexes, DT-C3d3 and SA-C3dg tetramers, revealed diverse effects on BCR-induced [Ca2+]i responses and humoral immunity. Fusing C3d3 to DT significantly reduced DT-specific IgM and IgG responses in mice compared with DT alone (Fig. 1, B–D) as previously described (22). However, DT-specific inhibition of immune responses by C3d was independent of CD21/35 expression (Fig. 1,D), and DT-C3d3 had minimal, if any, effects on BCR-induced [Ca2+]i responses in vitro (Fig. 3,A). These results contrast markedly with those obtained using SA and SA-C3dg as immunogens. C3dg primarily enhanced low-dose SA-specific IgM and IgG responses in wild-type mice, whereas wild-type mice immunized with high-dose SA or high-dose SA-C3dg generated similar Ab responses (Fig. 2,B). By comparison, CD21/35−/− mice produced significantly higher Ab responses to high- dose SA-C3dg (≥5-fold) than to high-dose SA or low-dose SA-C3dg (Fig. 2,C). This suggests that CD21 expression in wild-type mice may actually inhibit SA-specific Ab responses to high-dose SA-C3dg compared with low-dose SA-C3dg. In support of this interpretation, low SA-C3dg concentrations augmented BCR-induced [Ca2+]i responses in spleen B cells in vitro, whereas higher SA-C3dg concentrations inhibited BCR-induced [Ca2+]i responses (Fig. 3, B and C). Both the positive and negative effects of C3dg on BCR-induced [Ca2+]i responses required CD19 or CD21 expression (Fig. 3, D and E), ruling out the involvement of other C3dg receptors or regulatory signaling pathways in this process. Moreover, the concentrations of SA-C3dg that stimulated or inhibited BCR-induced [Ca2+]i responses were significantly below the maximal SA-C3dg binding capacity of B cells (Fig. 2,A). Thus, physiologically relevant concentrations of C3d-Ag complexes may have the potential to positively or negatively regulate B cell signaling and in vivo immune responses through both CD19/CD21 complex-dependent and independent pathways. The current results suggest that C3d may not function in vivo as a universal adjuvant, because the dose, nature, and composition of the C3d-containing complexes are likely to significantly influence humoral immunity. Moreover, CD21-dependent and CD21-independent pathways may individually influence how C3d modifies immune responses (Figs. 1 and 2). Thus, the finding that C3d can alter humoral immunity in the absence of CD21 expression does not imply that CD21 does also not contribute to this process when it is expressed. These factors may explain why Ab responses to some Ags are augmented in the presence of C3d (5, 6, 7, 8, 9, 10), whereas others are inhibited (20, 21, 22, 23). In addition, the use of different pathways may be influenced by structural differences and the valency of CD19/CD21 engagement, as in the cases of DT-C3d3 and SA-C3dg complexes. This may be crucial, because soluble DT-C3d3 and C3dg monomers alone did not induce [Ca2+]i responses or influence BCR-induced [Ca2+]i responses at any concentration tested (Fig. 3,G and data not shown). By contrast, HEL-C3d3 is structurally similar to DT-C3d3, but functions as a molecular adjuvant and boosts HEL-specific immune responses in vivo (5). However, in that model system, HEL alone induced [Ca2+]i responses in HEL-specific B cells, with augmented responses induced by HEL-C3d3. Because HEL monomers should not induce BCR cross-linking and [Ca2+]i responses, HEL in solution may have aggregated with the capacity to cross-link receptors and induce [Ca2+]i responses. These differences may explain why DT-C3d3 inhibited DT-specific immune responses, whereas HEL-C3d3 augmented HEL-specific immune responses. Because small differences in SA-C3dg concentrations either induced or inhibited BCR-induced [Ca2+]i responses (Fig. 3, B and C), relatively small differences in Ag-C3d concentrations administered or generated in vivo may have dramatic effects on the elicitation of humoral immunity. Thus, the molecular complexities of C3d-decorated Ags generated in vitro or in vivo are likely to preclude assumptions regarding the adjuvant or inhibitory properties of individual Ag-C3d complexes in vivo.
The current results also indicate that the innate sensing function of the CD19/CD21 complex can influence B cell signaling thresholds through CD19 regardless of BCR specificity (64). The CD19/CD21 complex is generally envisioned to serve as a costimulatory receptor where Ag-C3d complexes cross-link the CD19/CD21 and BCR complexes (5, 27, 28, 29, 53). However, the current study demonstrates that SA-C3dg ligation of the CD19/CD21 complex can independently augment BCR-induced [Ca2+]i responses when the two signals are given simultaneously (Fig. 3). Likewise, independent CD19 ligation by a dimeric IgA CD19 mAb modulated BCR-induced [Ca2+]i responses without co-cross-linking (Fig. 3 F), as previously described (44). Consistent with this, alterations in CD19 expression influence the phenotype and functions of all peripheral B cells, although only a small portion of B cells respond to Ags in vivo (32, 34, 35, 36). CD19 expression also influences B cell responses to mitogens that do not involve BCR signaling, further supporting a role for the CD19/CD21 complex as an intrinsic response regulator (32). Moreover, Ag-C3d complexes probably first encounter and bind to CD21 expressed on marginal zone or other CD21+ non-Ag-specific B cells in vivo and may not preferentially localize on rare Ag-specific B cells (13). In addition to regulating B cell signaling pathways through interacting with the CD21/CD19 complex, C3d has been reported to bind serum proteins and several surface proteins expressed on non-B cells, such as neutrophils, platelets, and monocytes (18, 65, 66, 67). A better understanding of these interactions may provide insight into the CD21-independent mechanisms by which C3d regulates humoral responses. A response regulatory role for the CD19/CD21 complex does not preclude the function of CD19/CD21 as a costimulatory receptor during B cell signaling, but expands the role of CD21 as an Ag-independent regulator of innate B cell signal transduction (64). Moreover, the finding that C3d can both positively and negatively regulate BCR signaling through the CD19/CD21 complex validates the physiological relevance of previous CD19 and CD21 mAb-based studies showing differential signaling roles for the CD19/CD21 complex (27, 28, 29, 37, 39, 40, 41, 42, 43, 44, 68, 69).
SA-C3dg amplification of [Ca2+]i responses through CD19/CD21 complex-dependent pathways (Fig. 3, D and E) provides a mechanistic explanation for the effects of SA-C3dg on B cell function. Binding of SA-C3dg to the CD19/CD21 complex significantly augmented and accelerated the kinetics of CD19 phosphorylation (Fig. 4,A). After its phosphorylation by Lyn, CD19 amplifies Src family PTK activity by recruiting additional Src family kinases that become phosphorylated by active Lyn bound to phospho-CD19 (51, 54). In fact, low-dose C3dg ligation of the CD19/CD21 complex significantly enhanced BCR-induced Lyn phosphorylation, Lyn/CD19 interactions, and the phosphorylation of other Src family kinases (Figs. 4,D and 6,A) that are amplified by CD19 (51, 54). SA-C3dg binding also amplified the BCR-induced phosphorylation of other signaling proteins downstream of CD19, including PLCγ, Akt, and p38 (Fig. 4,D). In previous studies by others, C3dg binding in a tetramer-like combination with very low concentrations of anti-IgM Ab (0.001 μg/ml) triggered B cell activation, but did not induce CD22 or SHP-1 phosphorylation (31). Likewise, low dose C3dg binding (1 μg/ml) did not augment or inhibit BCR-induced Syk phosphorylation (Fig. 6,C), consistent with previous observations that alterations in CD19 expression do not significantly affect Syk activation (54). By contrast, SA-C3dg binding significantly inhibited BCR-induced phosphorylation of the CD22/SHP-1 negative regulatory pathway (Fig. 7). CD19 ligation by mAbs has been previously shown to augment [Ca2+]i responses by inhibiting CD22 phosphorylation (44). Thus, low dose CD21 ligation by SA-C3dg activates established CD19/CD21 complex signaling pathways that synergize with BCR-induced signaling pathways to augment [Ca2+]i responses.
High-valency CD19/CD21 complex ligation by SA-C3dg (50 μg/ml) is likely to inhibit BCR-induced [Ca2+]i responses by inhibiting CD19, Lyn and Syk phosphorylation. High-dose SA-C3dg binding reduced BCR-induced CD19 phosphorylation (Fig. 4,A) and Lyn/CD19 associations (Fig. 6,A) to levels induced by BCR ligation alone. High-dose SA-C3dg binding also reduced the phosphorylation of Lyn and other Src family kinases, PLCγ1, Akt, and p38, to levels similar to those seen with BCR ligation alone (Fig. 4,D). Interestingly, high-dose CD19/CD21 ligation almost completely eliminated BCR-induced Syk phosphorylation (Fig. 6,B), an important mediator of BCR-induced protein phosphorylation and [Ca2+]i responses (70). CD21 ligation with high concentrations of mAb has been shown to recruit Lyn away from the BCR complex (37), which would negatively affect BCR-induced Syk kinase activity (57, 58). High doses of a multimeric CD19 mAb had similar effects on BCR-induced [Ca2+]i responses and the phosphorylation of downstream signaling molecules (Figs. 3–6). The effects of CD19 ligation by mAb were more dramatic than those of SA-C3dg in most cases, probably due to the fact that CD19 exists in molar excess of CD21 on the surface of B cells (24, 53, 71). High-level CD19/CD21 ligation also reduced CD22 phosphorylation to levels below those induced by simultaneous low-level CD19/CD21 and BCR ligation (Fig. 7). Thus, excessive CD19/CD21 complex ligation is likely to primarily inhibit BCR-induced [Ca2+]i responses through mechanisms that inhibit Lyn and Syk kinase activities. Thus, CD19/CD21 exerts its regulatory role by altering the phosphorylation of both positive and negative regulatory molecules.
In summary, these results demonstrate that C3d and C3dg can negatively regulate BCR-induced [Ca2+]i responses and Ag-specific immune responses depending on the extent of CD19/CD21 complex ligation and the nature of the Ag-C3d complexes. Although some Ag-C3dg complexes are immunostimulatory, complexes containing multiple copies or high levels of C3d may inhibit B cell activation and dampen immune responses. Because C3dg is often a component of immune complexes, this may be an autoregulatory feature of the complement system that reduces humoral responses in the face of ongoing immune complex disease. The concept of CD21 as both a positive and a negative regulator of B cell signaling also provides a mechanistic explanation for why CD21 deficiency or defective CD21 ligand binding contributes to the development of autoimmunity in lpr/lpr and (NZB × NZW)F1 mice (45, 46, 47), whereas CD21−/− mice are generally hyporesponsive to Ag challenge. Thus, negative regulation of B cell activation by the CD19/CD21 complex may be important for balancing immune responses.
We thank Dr. Clifford Lowell for providing Lyn−/− mice for these studies.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants CA96547, CA105001, and AI56363. K.M.H. is a Special Fellow of the Leukemia and Lymphoma Society.
Abbreviations used in this paper: HEL, hen egg lysozyme; [Ca2+]i, intracellular calcium concentration; DT, nontoxic B fragment of diphtheria toxin; PLCγ1, phospholipase Cγ1; pTyr, phosphotyrosine; SA, streptavidin; SHP-1, Src homology 2 domain-containing protein tyrosine phosphatase-1.