Recent studies demonstrate that MHC class II molecules can signal via associated Ig-αβ dimers, signal transducers previously thought to function only in B cell Ag receptor (BCR) signaling. Surprisingly, the biologic outputs of MHC class II and BCR ligation (by thymus-dependent Ags) differ, e.g., MHC class II signaling leads to robust proliferation and extension of pseudopods. It seemed possible that these differences might be due, at least in part, to differential use of inhibitory coreceptors thought to modulate membrane Ig signals. In this study, we demonstrate that CD22, an inhibitory BCR coreceptor, neither associates with nor functions in MHC class II/Ig-αβ signaling. Interestingly, CD22 is actively excluded from cell surface MHC class II aggregates.
During cognate T cell-B cell interactions, multiple receptor-ligand pairs become localized at the cell contact interface. Many of these proteins become further ordered into subregions, termed central and peripheral supramolecular activation clusters (SMACs)3 (1, 2). This organization increases the local concentration of signaling molecules and ensures that certain effectors are sequestered away from some proteins and colocalized with others. Although T cell SMAC formation can occur in the absence of APCs (2), it is currently unknown whether transmembrane signaling in one or both cells influences the formation, stabilization, and movement of molecules into and out of SMACs in vivo. However, although T-B conjugate formation does not depend on the T cell recognizing its specific Ag, the formation of SMACs at cell contact points is seen only upon Ag recognition (1, 2). Thus, Ag-driven aggregation of both MHC class II and TCR may be important for productive T-B collaboration.
Aggregation of MHC class II on resting B cells leads to production of cAMP and activation of protein kinase C (3, 4, 5). However, in B cells activated by Ag and/or IL-4, aggregation of MHC class II induces intracellular Ca2+ concentration ([Ca2+]i) mobilization, tyrosine kinase activation, dendritic extensions, and proliferation (6, 7, 8). Because B cells actively engulf TCR-coated beads (9), it is possible that the formation of pseudopods by B cells following MHC class II ligation in vitro reflects molecular events important for in vivo dynamics of T-B conjugates. Triggering of these primed cell responses is independent of the cytoplasmic domains of MHC class II (3). Recent studies demonstrate that signaling by MHC class II in primed cells occurs by association of the B cell Ag receptor (BCR) signal-transducing substructure, Ig-α/Ig-β dimers, with MHC class II, and that transmembrane signaling by MHC class II is induced upon cognate TCR interactions (9). Extensive mutational analyses have determined that the ability of MHC class II to associate with Ig-αβ is required for I-A-mediated calcium mobilization and induction of whole-cell protein tyrosine phosphorylation in primed B cells.4 Although aggregation of both BCR and MHC class II on primed cells induces [Ca2+]i mobilization and tyrosine kinase activation, the responses differ at the biologic level. Although both MHC class II aggregation and BCR ligation (by thymus-dependent Ags) leads to increased expression of activation markers, only MHC class II signals cause extension of dendrite-like pseudopodia (6).
We hypothesized that differences in MHC class II and BCR signaling may be due to differential use of coreceptors. Signal transduction by the BCR can involve several coreceptors, including CD19/21, CD22, FcγRIIB, and CD72. The coreceptors involved in MHC class II signal transduction are not well characterized.
CD22 is a 140-kDa sialic acid-binding Ig-like lectin superfamily member that binds α2,6-linked sialic acid residues (in particular amino acid contexts) and is tyrosyl phosphorylated upon BCR aggregation (reviewed in Ref. 10). CD22 contains three immunoreceptor tyrosine-based inhibitory motifs in its cytoplasmic domain. Mice deficient in CD22 display increased sensitivity to BCR aggregation (measured by [Ca2+]i mobilization and proliferation), but no noticeable differences in some thymus-dependent immune responses (11, 12). Although some studies suggest that CD22 may interact with sialic acid associated with membrane (m)Ig, and thereby modulate BCR signaling selectively, these findings are controversial. The potential interaction of MHC class II with CD22 has not been explored. CD22 is phosphorylated by the Src family kinase Lyn and exerts its negative effects in part by recruiting the tyrosine phosphatase Src homology 2 domain-containing phosphatase-1 (SHP-1) to signaling complexes. This inhibitory signaling loop sets thresholds for autoimmunity in mice (13). SHP-1 substrates are not well defined, but are proposed to include CD22, CD72, Vav, B cell linker protein, Syk, Ig-α, and Ig-β (14, 15).
In this report, we addressed whether the inhibitory BCR coreceptor CD22 modulates MHC class II signal transduction, and whether CD22 usage can account for any of the signaling differences observed following BCR and MHC class II aggregation. We report that CD22 becomes tyrosyl phosphorylated upon BCR but not MHC class II aggregation. Consistent with these findings, we show that the CD22 effector SHP-1 becomes tyrosyl phosphorylated in response to BCR but not MHC class II aggregation. Studies using knockout mice indicate that, although CD22 negatively regulates BCR-mediated [Ca2+]i mobilization and proliferation, it does not affect MHC class II-induced responses. We were able to detect CD22 enrichment in IgM but not MHC class II immunoprecipitates. Interestingly, although CD22 was recruited to mIgM aggregates, it was actively excluded from MHC class II aggregates on stimulated cells. Finally, forced coaggregation of CD22 and MHC class II induced Ca2+ mobilization responses similar to those induced by BCR aggregation, and inhibited formation of pseudopods.
Materials and Methods
Cell preparation and culture
Splenic B cells from Ig-transgenic 3-83, C57BL6, and CD22−/− mice were used in these experiments. Spleens were excised from 2- to 4-mo-old mice; single-cell suspensions were prepared, depleted of erythrocytes using Gey’s solution, and washed once in IMDM. T cells were lysed using anti-Thy1 mAbs and guinea pig complement. B cells were isolated by Percoll gradient centrifugation, washed with IMDM, and cultured at 2–3 × 106 cells/ml (7% CO2; 37°C) in IMDM containing 100 U/ml penicillin, 100 μg/ml streptomycin, 50 μg/ml gentamicin, 2 mM l-glutamine, 1 mM sodium pyruvate, 50 μM 2-ME, and 10% FCS (HyClone Laboratories, Logan, UT). Priming of MHC class II signal transduction was accomplished by culturing cells for 12–18 h in 50 U/ml recombinant murine IL-4. For pseudopod formation assays, both IL-4 and 10 μg/ml low-affinity anti-IgM were used to prime B cells. All samples were incubated for 15 min with anti-Fcγ Abs before stimulation to block IgG binding to these receptors. For plate-bound stimulations, tissue culture plates were coated with anti-MHC class II, anti-CD22, and anti-H-2K Abs (diluted in PBS) for 18 h at 37°C. Wells were washed three times with PBS to remove nonbound Ab molecules.
Abs and reagents
The following mAbs were purified from hybridoma supernatant and used for stimulation, immunoprecipitation, and staining experiments: anti-I-Ad/b (D3.137), anti-CD22 (Cy34), low-affinity anti-μ H chain (Bet-2), high affinity anti-μ (b-7-6), anti-FcγRII/III (2.4G2), and anti-H-2Kd/b/k (M22.214.171.124.8). Purified polyclonal rabbit anti-SHP-1 used for immunoprecipitation and immunoblotting was from Upstate Biotechnology (Lake Placid, NY). Anti-CD19 and anti-CD22 cytoplasmic tail sera used for immunoblotting were raised in New Zealand White rabbits. Anti-MHC class II (I-Aα) for immunoblotting was kindly provided by Dr. I. Mellman (Yale University, New Haven, CT). Additional reagents used for immunoblotting and staining were as follows: donkey anti-mouse IgG1-HRP (Zymed, San Francisco, CA), anti-mouse IgM-HRP (Zymed), protein A-HRP (Zymed), Cy3-streptavidin (Caltag, Burlingame, CA), Cy3-goat anti-mouse IgG1 (Caltag), Cy5-goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA), and Cy5-donkey anti-rat IgG (Jackson ImmunoResearch Laboratories). rIL-4 was purified from J558L-IL-4 plasmacytoma supernatant. Protein A- and streptavidin-Sepharose used for immunoprecipitation were from Amersham Pharmacia Biotech (Uppsala, Sweden).
Analysis of [Ca2+]i
Cells were resuspended at 5 × 106/ml in IMDM (2.5% FCS) and incubated for 45 min at 37°C in 5 μM Indo 1-AM (Molecular Probes, Eugene, OR). For aggregation of MHC class II, cells were incubated with biotinylated anti-MHC class II (1 μg/ml) for 15 min at room temperature. For coaggregation of MHC class II and CD22 or MHC class I, cells were incubated with biotinylated anti-MHC class II Abs, as described above, and biotinylated anti-MHC class I or biotinylated anti-CD22. Doses of biotinylated anti-MHC class I and biotinylated anti-CD22 were selected, which yielded comparable staining by flow cytometry. Cells were washed with IMDM and resuspended at 1.5 × 106/ml for analysis on an LSR flow cytometer (BD Biosciences, Mountain View, CA). Signaling was triggered by addition of avidin and anti-μ as indicated in the figures. Data analysis was performed using FlowJo software (Tree Star, San Carlos, CA).
For analysis of Ca2+ release from intracellular stores, 0.5 M EGTA was added 1 min before stimulation to buffer the free [Ca2+] in the medium (extracellular Ca2+ concentration ([Ca2+]o)) to 60 nM (isotonic with respect to the cell cytoplasm). Four minutes poststimulation, 1 M CaCl2 was added to adjust [Ca2+]o to 1.3 mM, allowing analysis of Ca2+ influx.
Immunoprecipitation and immunoblotting
IL-4-primed splenic B cells were resuspended at 20 × 106/ml in IMDM (10% FCS). Cells were incubated with biotinylated anti-I-Ab/d at 20 μg/ml for 20 min at room temperature. Samples were washed once and resuspended at 20 × 106 cells/ml in IMDM. Stimulations were performed for 5 min at 37°C using 5 μg/ml anti-μ or 20 μg/ml avidin. Samples were lysed for 10 min on ice in 1 ml 0.5% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate lysis buffer (10 mM NaF, 0.4 mM EDTA, 2 mM sodium orthovanadate, 10 mM tetrasodium pyrophosphate, 1 mM PMSF, 1 mM aprotinin, 1 mM α-1-antitrypsin, and 1 mM leupeptin) followed by centrifugation at 14,000 rpm for 10 min in an Eppendorf centrifuge. CD22 was immunoprecipitated using directly coupled anti-CD22 Ab-conjugated Sepharose beads. CD19 and SHP-1 were immunoprecipitated with 10 μl of rabbit serum or purified IgG prebound to 20 μl of protein A-Sepharose beads. Immunoprecipitations were performed for 1 h (or overnight for Fig. 5 B) at 4°C with constant mixing. Beads were pelleted by brief centrifugation and washed three times with 1 ml of lysis buffer. Samples were boiled in reducing sample buffer and fractionated by 10% SDS-PAGE. Proteins were transferred to polyvinylidene difluoride membranes and visualized using specific Abs followed by ECL (NEN, Boston, MA). Relative induction of CD22 phosphorylation was calculated by obtaining a ratio of the intensity of the anti-phosphotyrosine signal to the anti-CD22 immunoblot signal. These ratios were then divided by the ratio obtained for unstimulated samples to obtain relative induction. Band intensities were calculated using NIH Image.
Primed B cells were incubated with biotinylated anti-MHC class II for 20 min at room temperature and washed once with IMDM. Samples were stimulated with either polyclonal rabbit Cy5-anti-IgM, FITC-streptavidin, or Cy3-streptividin for 30 min at 37°C. Cells were placed on poly-l-lysine-coated coverslips for 2 min at 37°C and fixed with 4% paraformaldehyde, followed by permeabilization for 5 min with 0.2% Triton X-100. Coverslips were blocked with PBS (2.5% FBS) for 5 min and probed with Cy5-anti-IgM, biotinylated anti-MHC class II, anti-CD22, anti-CD19, or anti-MHC class I Abs for 1 h at room temperature. Samples were washed several times in PBS and probed with appropriate secondary reagents for 1 h at room temperature. After washing several times, coverslips were mounted on microscope slides with mounting solution (2 mg/ml o-phenylenediamine in 90% glycerol). Images were captured and analyzed using a Leica (Deerfield, IL) DMRXA microscope, a SensicamQE camera (Cooke, Auburn Hills, MI), and Slidebook imaging software (Intelligent Imaging Innovations, Denver, CO).
Splenic B cells from wild-type and CD22−/− mice were negatively selected using anti-CD43-coated microspheres (Miltenyi Biotec, Auburn, CA) and magnetic separation according to the manufacturer’s recommendations. Cells were determined by flow cytometry to be >95% B220+ (data not shown). Cells were incubated at 5 × 106/ml in HBSS with 2 μM CFSE (Molecular Probes) for 5 min at 22°C. Cells were then washed with HBSS and cultured for 18 h at 2 × 106/ml with low-affinity anti-μ (Bet-2; 2.5 μg/ml) and IL-4 (50 U/ml) in IMDM (10% FCS, 2% nonessential amino acids, and 50 μM 2-ME). Samples were then transferred to noncoated or anti-I-A-coated 24-well plates and cultured for 72 h for 3 days. Proliferation was analyzed using a FACScan cytometer (BD Biosciences) and FlowJo analysis software (Tree Star). Data are presented as percentage of maximum population.
Results and Discussion
Aggregation of MHC class II and IgM induces dissimilar Ca2+ influx in primed splenic B cells
To explore molecular mechanisms underlying differences between BCR and MHC class II signaling, we compared Ca2+ mobilization responses to BCR and MHC class II aggregation in purified splenic B cells. We found that aggregation of MHC class II induces more robust Ca2+ mobilization than aggregation of BCR in primed cells at multiple stimulatory doses (Fig. 1 A). At its maximal stimulatory dose, anti-μ-induced calcium mobilization did not approach that induced by anti-MHC class II.
To investigate the basis of the dissimilar Ca2+ mobilization seen in response to MHC class II and IgM aggregation, we analyzed the intracellular release and influx components of the response separately following maximal stimulation with anti-μ or anti-I-A Abs. We observed greater Ca2+ influx in response to MHC class II aggregation compared with BCR aggregation, but comparable release from intracellular stores (Fig. 1 B). These results suggest two possibilities. First, a post-intracellular Ca2+ release mechanism may selectively limit the BCR response or enhance the MHC class II response. Alternatively, this mechanism may not be influx targeted, but merely late acting.
The different intensities of MHC class II- and mIgM-mediated calcium mobilization observed in Fig. 1 A may reflect different degrees of cell surface receptor aggregation (resulting from the use of different ligands). However, the response magnitude differences are not entirely explained by Ab and/or aggregate dissimilarity, because under stimulatory conditions used in these experiments, the entire cell surface pool of each receptor is recruited into a single membrane aggregate (data not shown). Additionally, this difference in magnitude was seen when multiple anti-Ig and anti-I-A Abs were compared (data not shown) and may indicate that MHC class II has a greater intrinsic ability to transduce signals leading to Ca2+ mobilization.
MHC class II-induced Ca2+ mobilization is normal in CD22−/− B cells
It has been shown previously that CD22 limits extracellular Ca2+ influx mediated by BCR ligation (16). Based on these observations, and those shown in Fig. 1, we hypothesized that, in contrast to the BCR, MHC class II signal transduction is not regulated by CD22. To address this hypothesis, we compared [Ca2+]i following MHC class II and BCR ligation in CD22−/− and wild-type B cells. We observed modestly elevated basal [Ca2+]i in resting CD22−/− cells, perhaps implying that these cells experience more robust tonic signals. As reported by others, CD22−/− B cells displayed enhanced Ca2+ influx upon BCR aggregation (Fig. 2,A). The [Ca2+]i mobilization response to BCR aggregation in CD22−/− B cells was similar to the response of wild-type cells to MHC class II aggregation. In contrast to BCR stimulation, MHC class II-induced Ca2+ mobilization was similar in CD22−/− and wild-type B cells (Fig. 2,A). These findings were not a result of maximal stimulation, because we obtained similar results at suboptimal stimulatory doses (Fig. 2 B). These results suggested that CD22 is functionally linked to BCR, but not MHC class II signal transduction pathways.
CD22 is tyrosyl phosphorylated in response to BCR but not MHC class II aggregation
To more directly address the hypothesis that CD22 does not negatively regulate MHC class II signal transduction, we analyzed tyrosyl phosphorylation of CD22 and CD19 (a positive regulator of BCR signaling) in response to receptor aggregation in primed splenic B cells. Although CD19 was tyrosyl phosphorylated in response to both BCR and MHC class II aggregation, CD22 was only phosphorylated upon BCR aggregation (Fig. 3 A). This finding indicates that the MHC class II and BCR signaling complexes share some accessory molecules, including CD19, and use others uniquely, such as CD22.
These results appear inconsistent with recently published findings of Bobbitt et al. (17), who report that both CD19 and CD22 are tyrosyl phosphorylated following MHC class II aggregation. One explanation is that Bobbitt et al. (17) used a B lymphoma to biochemically study coreceptor cooperation with MHC class II, whereas the studies presented here used primary B cells. Thus, it is possible that the degree to which CD22 modulates MHC class II signaling differs between ex vivo B cells and this lymphoma.
The findings presented above predict that downstream signaling events influenced by CD22 cooperation would be different upon MHC class II and BCR aggregation. To address this possibility, we analyzed SHP-1 tyrosyl phosphorylation upon aggregation of either MHC class II or the BCR. As reported by others, we observed increased tyrosyl phosphorylation of SHP-1 upon BCR ligation (Fig. 3 B). Although tyrosyl phosphorylation of SHP-1 was induced upon MHC class II aggregation, it was considerably less robust, consistent with the finding that CD22 is not involved in MHC class II signal transduction.
CD22 coimmunoprecipitates with IgM, but not MHC class II, and is excluded from MHC class II aggregates
As discussed previously, CD22 associates directly with the BCR signaling complex (18). The data described above suggested that CD22 may not associate physically with the MHC class II signal transduction complex. To address this possibility, we immunoblotted IgM and MHC class II immunoprecipitates using anti-CD22 Abs. Additionally, we used fluorescence microscopy to assess the spatial localization of CD22, plasma membrane IgM (mIgM), and MHC class II in nonstimulated and stimulated B cells.
We found that CD22 was present in anti-μ immunoprecipitates from B cells before and following aggregation of mIgM (Fig. 4,A). Conversely, we were unable to detect CD22 in MHC class II immunoprecipitates either before or after aggregation of I-A. However, in agreement with previously reported findings, we found that nearly all of the cell surface CD22 colocalized with mIgM following BCR aggregation (19). In contrast, CD22 was actively excluded from I-A aggregates formed following MHC class II ligation (Fig. 4,B). In control experiments, the localization of MHC class I was unaffected by MHC class II aggregation (Fig. 4,C). Additionally, a portion of the plasma membrane CD19 was enriched in MHC class II aggregates (Fig. 4 C). These results demonstrate that CD22 is a constitutive member of the BCR, but not the MHC class II signal transduction complex. Conversely, CD19 likely exists on the cell surface in distinct pools, associated with receptors such as MHC class II, mIgM, and CD21 (17, 20, 21, 22).
In previous studies (9), we found that the cell surface localization of CD22 on the primed lymphoma line K46 was not altered following stimulation with TCR-coated microspheres. These data corroborate biochemical evidence presented here in Figs. 2, 3, and 6 but do not reveal the spatial segregation of CD22 and I-A observed in Fig. 4 B. Although we cannot exclude the possibility that this is due to differences in cell systems used (ex vivo B cells vs K46), it is possible that only stimulation with relatively high-affinity ligands induces the spatial segregation of CD22 and MHC class II. Alternatively, because D3.137 Abs bind to all cell surface I-Ab/d molecules but DO.11.10 TCR monomers ligate only those I-Ad molecules presenting the OVA peptide, the exclusion of CD22 from MHC class II aggregates may be detectable only above a critical MHC class II-peptide density threshold. Finally, such spatial segregation of CD22 from MHC class II may occur only if the MHC ligand is laterally mobile. It will be important to determine the molecular basis for the spatial segregation of MHC class II and CD22.
Forced coaggregation of CD22 and MHC class II reduces Ca2+ influx, leads to CD22 tyrosyl phosphorylation, and inhibits pseudopod formation
As an additional challenge to the hypothesis that CD22 differentially regulates Ig-αβ signals mediated by BCR and MHC class II, we compared Ca2+ mobilization responses of primed B cells to MHC class II aggregation and forced MHC class II/CD22 coaggregation. We found that coaggregation of MHC class II and CD22 diminishes the Ca2+ influx response compared with aggregation of MHC class II alone (Fig. 5,A). Analogous inhibition was not observed upon independent aggregation of I-A and CD22 (data not shown). In specificity controls, coaggregation of MHC class I with MHC class II also did not significantly alter the response. The [Ca2+]i mobilization response to coaggregation of MHC class II and CD22 was similar to the [Ca2+]i mobilization response to BCR aggregation. To address whether CD22-mediated inhibition of MHC class II signals in this experiment was due to inhibitory signaling, we analyzed CD22 phosphorylation following coaggregation with MHC class II. We observed induced phosphorylation of CD22 upon forced coaggregation with MHC class II (Fig. 5 B), suggesting that MHC class II is capable of recruiting the appropriate kinases for initiation of CD22-mediated inhibitory signaling.
To address the biologic significance of differential CD22 usage by the BCR and MHC class II, we assessed the effect of forced CD22 coaggregation on the ability of B cells to form pseudopods upon ligation of MHC class II. We analyzed pseudopod formation by primed splenic B cells cultured on plates coated with anti-MHC class II Abs alone, or anti-MHC class II and anti-CD22 Abs. As shown in Fig. 5, C and D, we observed dose-dependent inhibition of pseudopod formation by forced CD22/MHC class II coaggregation. Specificity controls illustrated that pseudopod formation is not inhibited by coaggregation of MHC class I and MHC class II. As expected, BCR aggregation did not induce pseudopod formation. These data demonstrate that CD22 does not regulate MHC class II signal transduction under normal circumstances, but can be forced to do so.
If BCR and I-A signaling via Ig-αβ differs only in usage of CD22, one would predict that CD22-deficient B cells would form pseudopods on anti-IgM-coated plates. Interestingly, CD22−/− B cells did not change their morphology following IgM aggregation (data not shown). In agreement with data presented in Figs. 2 and 6, CD22-deficient B cells formed pseudopodic extensions normally following MHC class II aggregation, even at suboptimal stimulatory doses (data not shown). Because forced CD22 coaggregation can inhibit MHC class II-mediated pseudopod formation, we conclude that lack of CD22 inhibitory signals is important for these morphologic transitions. However, because CD22 deficiency is not sufficient to convey this phenotype following BCR aggregation, other qualitative differences must exist between MHC class II and BCR signal transduction that mediate changes in cell morphology.
B cells proliferate comparably following MHC class II and BCR aggregation in the absence of CD22 function
To further explore the role of CD22 in biologic responses of B cells, we analyzed proliferation of wild-type or CD22−/− B cells following mIgM and MHC class II ligation. Purified splenic B cells were labeled with CFSE, primed with IL-4 and low-affinity anti-IgM (to mimic a thymus-dependent Ag), and cultured on noncoated or anti-I-A-coated tissue culture plates. As reported by others (11, 12), we found that CD22-deficient B cells are more sensitive to anti-μ-induced proliferation (Fig. 6). However, the proliferation of wild-type and CD22-deficient B cells is comparable following MHC class II ligation (Fig. 6). The modest increases in MHC class II-mediated proliferation of B cells from CD22−/− mice may reflect enhanced priming of MHC class II signaling by low affinity anti-μ. Most importantly, proliferative responses of CD22-deficient B cells following mIgM aggregation approach the levels seen following I-A aggregation on wild-type B cells. These data extend our previous findings (Figs. 2–4) by demonstrating that differential use of CD22 by I-A and mIgM has both biochemical and biologic consequences for B cell responses. Additionally, the data demonstrate that negative regulation by CD22 can fully account for some differences in the consequences of MHC class II and mIgM signal transduction (Ca2+ mobilization and proliferation) but not others (pseudopod formation).
Finally, although IL-4 is sufficient to prime B cells for MHC class II-induced calcium mobilization, optimal priming of MHC class II-mediated pseudopod formation and proliferation occurs following combined stimulation with IL-4 and anti-IgM (data not shown and Ref. 6). The reasons underlying these differential requirements are currently unclear.
The studies described in this report were undertaken to determine how BCR and MHC class II transduce signals with distinct biologic consequences despite their use of the same transmembrane transducers, Ig-α and Ig-β. Our experiments suggest that BCR- mediated Ca2+ influx is constrained relative to that induced through MHC class II. Studies using knockout mice indicated that CD22 does not modulate MHC class II-mediated responses. We further show that CD22 becomes tyrosyl phosphorylated upon BCR but not MHC class II aggregation, and that this correlates with tyrosyl phosphorylation of the CD22 effector SHP-1. In addition, we provide evidence that CD22 is associated with BCR aggregates, but is actively excluded from MHC class II aggregates. Importantly, forced coaggregation of MHC class II and CD22 resulted in a Ca2+ mobilization response similar to that induced by BCR ligation, and inhibited pseudopod formation upon MHC class II aggregation. Finally, CD22 deficiency had little effect on the ability of B cells to proliferate following MHC class II ligation. Most significantly, differential CD22 usage, at least in part, does underlie differences in proliferation to BCR and MHC class II aggregation.
The observation that CD22 associates with mIgM but not MHC class II may be explained by recent studies which demonstrate that the lectin-binding domain of CD22 is required for participation in BCR signaling (23, 24). Thus, it is possible that CD22 associates with mIgM via interactions with sialic acid moieties that are not present in MHC class II extracellular domains, or are present in an unfavorable amino acid context. The finding that CD22 is actively excluded from MHC class II aggregates indicates that CD22 function is subject to additional levels of regulation beyond the presence or absence of appropriate sialic acid ligands. This is corroborated by recent evidence that the longer cytoplasmic domains of mIgG disallow association of mIg with CD22 in B cell lymphoma models (25).
It is perhaps not surprising that mechanisms have evolved to enhance signal output through MHC class II. In physiologic responses, TCR-MHC class II interactions are typically of low affinity. Furthermore, only a very small proportion of MHC class II on an APC is associated with peptide recognized by a single T cell. Thus, a limited number of MHC class II molecules are recognized by TCR. It is noteworthy that recent evidence indicates many MHC class II molecules bearing irrelevant peptides are recruited to T-B contact regions (26). However, it is not known whether these noncognate interactions lead to MHC class II signaling.
We report here the novel observation that unique qualities of ligand-binding substructures (e.g., mIgM or MHC class II) determine, by secondary associations with accessory proteins, the quality of signals transduced by primary signal transducing substructures (e.g., Ig-αβ). Our studies underscore the importance of receptor-associated modulators in diversification of biochemical and biologic responses that follow receptor ligation.
Current evidence indicates that Ig-αβ dimers play dual roles during B cell activation by transducing signals following Ag encounter and during cognate T-B collaboration. The findings reported here suggest that CD22 limits Ig-αβ signals and subsequent B cell activation following Ag encounter, but once vetted, maximal Ig-αβ signaling occurs upon TCR binding. Thus, Ig-αβ signal strength/quality may be limited during the early phases of thymus-dependent B cell responses to circumvent activation of potentially harmful autoreactive cells.
We thank Drs. Joseph Dal Porto and Stephen Gauld for critical reading of this manuscript.
These studies were supported by National Institutes of Health Grant AI20519. J.C.C. is an Ida and Cecil Green Professor of Immunology.
Abbreviations used in this paper: SMAC, supramolecular activation cluster; [Ca2+]i, intracellular Ca2+ concentration; [Ca2+]o, extracellular Ca2+ concentration; BCR, B cell Ag receptor; m, membrane; SHP-1, Src homology 2 domain-containing phosphatase-1.
J. Stolpa, D. Mills, and J. Cambier. The connecting peptide region of the α chain of MHC class II molecules is required for association with the signal-transducing heterodimer Ig-α/β. Submitted for publication.