Protein-tyrosine Phosphatase SHP-1 Is Dispensable for FcγRIIB-mediated Inhibition of B Cell Antigen Receptor Activation*

The inhibitory Fc receptor, FcγRIIB, provides a signal that aborts B cell antigen receptor activation, blocking extracellular calcium influx. Because the protein-tyrosine phosphatase SHP-1 binds tyrosyl phosphorylated FcγRIIB and FcγRIIB-mediated inhibition is defective in motheaten (me/me) mice, which do not express SHP-1, it was proposed that SHP-1 mediates FcγRIIB signaling in B cells (D’Ambrosio, D., Hippen, K. L., Minskoff, S. A., Mellman, I., Pani, G., Siminovitch, K. A., and Cambier, J. C. (1995) Science 268, 293–297). However, SHP-1 is dispensable for FcγRIIB-mediated inhibition of FcεRI signaling in mast cells (Ono, M., Bolland, S., Tempst, P., and Ravetch, J. V. (1996) Nature 383, 263–266), prompting us to re-examine the role of SHP-1 in FcγRIIB signaling in B cells. We generated immortalized sIgM+, FcγRIIB+ cell lines fromme/me mice and normal littermates. Co-ligation of FcγRIIB and the sIgM antigen receptor inhibits calcium influx in both cell lines. Inhibition is reversed by preincubation with anti-FcγRIIB antibodies, indicating that it is mediated by FcγRIIB. The inositol 5′ phosphatase SHIP is recruited to tyrosyl-phosphorylated FcγRIIB in both cell lines. FcγRIIB-mediated CD19 dephosphorylation also occurs in the presence or the absence of SHP-1. Our results establish that SHP-1 is dispensable for FcγRIIB-mediated inhibition of sIgM antigen receptor signaling.

either enhance or dampen the BCR signal. These initial events lead to Ras activation, PI3-K activation, phosphoinositide turnover, and an increase in intracellular free calcium (reviewed in Ref. 1).
By co-ligating the BCR and the inhibitory Fc receptor, Fc␥RIIB, immune complexes activate a negative feedback mechanism that inhibits antibody production (reviewed in Refs. 2 and 3). Fc␥RIIB-mediated inhibition in B lymphocytes can be studied experimentally by comparing responses evoked by cross-linking the BCR with F(abЈ) 2 anti-Mu, which engages the BCR alone, with those initiated with intact anti-Mu Ig, which co-ligates Fc␥RIIB and the BCR. Co-ligation results in inhibition of BCR-evoked inositol triphosphate production and extracellular calcium influx (4,5). Sustained calcium influx is required for multiple cellular processes, including proliferation and differentiation (reviewed in Refs. 6 and 7). Low concentrations of calcium ionophores override Fc␥RIIB-mediated inhibition of A20 B cell activation, indicating that inhibition of calcium influx is a (the) key mechanism by which antigenmediated signals are aborted by Fc␥RIIB signals (4). An analogous mechanism operates in mast cells, where co-engagement of Fc␥RIIB and Fc⑀RI results in inhibition of Fc⑀RIgenerated responses (9 -11).
A single tyrosyl residue, situated within the 13-amino acid ITIM (immune receptor tyrosine-based inhibitory motif) in the cytoplasmic domain of Fc␥RIIB, is required for Fc␥RIIB-mediated inhibition of BCR signaling (12). SHP-1, an SH2-domain containing protein-tyrosine phosphatase (PTP), binds to the tyrosyl-phosphorylated Fc␥RIIB ITIM motif both in vitro and in vivo (13). Because B cells from motheaten (me/me) and motheaten viable (me v /me v ) mice, which have defective SHP-1 alleles, failed to demonstrate Fc␥RIIB-mediated inhibition, it was proposed that the Fc␥RIIB inhibitory signal is mediated by SHP-1 (13,14). Consistent with this model, a recent report demonstrated that CD19 is dephosphorylated upon co-crosslinking of Fc␥RIIB and the BCR, leading to decreased binding of the p85 subunit of PI-3K to CD19 (15). This led to the proposal that SHP-1 regulates intracellular calcium levels by controlling CD19 tyrosyl phosphorylation, which in turn controls associated PI-3K activity (15). Interestingly, me/me mice exhibit a plethora of immune system defects, including hypergammaglobulinemia and a high level of circulating immune complexes (reviewed in Ref. 16). Conceivably, one or more of these disorders might be explained by defective Fc␥RIIB signaling.
However, the role of SHP-1 in Fc␥RIIB-mediated inhibitory signaling has been a subject of controversy. Mice lacking Fc␥RIIB do not exhibit uncontrolled antibody production when challenged with antigen (9). Moreover, as revealed by studies of mast cells from me/me mice, SHP-1 clearly is dispensable for * This work was supported by National Institutes of Health Grants CA66600 (to B. G. N.) and AR43773 and AI15803 (to H. H. W.) and National Research Service Award Grant CA65048 (to M. J. S. N.). 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.
Fc␥RIIB-mediated inhibition of Fc⑀RI signaling (17). Ono et al. have proposed that the inositol phosphatase SHIP (SH2 domain containing 5Ј-inositol phosphatase), which also binds to the Fc␥RIIB ITIM, is the key mediator of Fc␥RIIB signaling (17). The situation is complicated further by a report that the other mammalian SH2-containing PTP, SHP-2, also binds to this ITIM, at least in vitro (14). Although it is possible that Fc␥RIIB employs unique signaling mechanisms in different cell types, these findings have raised questions as to whether SHP-1 actually is required for Fc␥RIIB-mediated inhibition in B cells.
To develop a model system for studying the role of SHP-1 in multiple B cell signaling pathways, we generated B cell lines from me/me (ME) and normal (N) littermate mice by immortalization with the retrovirus J2 (18,19). N and ME cell lines express similar surface markers and respond to BCR stimulation and Fc␥RIIB/BCR co-ligation. We find that both cell lines exhibit comparably decreased calcium influx when Fc␥RIIB is co-cross-linked with the BCR. Consistent with previous reports, CD19 tyrosyl phosphorylation decreases upon co-crosslinking of Fc␥RIIB and the BCR in N cells. However, dephosphorylation also is observed in the ME line. Our results establish that SHP-1 is dispensable for both biochemical and biological responses to Fc␥RIIB engagement in B cells.

EXPERIMENTAL PROCEDURES
Generation of Immortalized me/me and Normal B Cells-J2 producer cells (20) were obtained from Dr. Christopher Myers of the University of Tennessee, Memphis, TN. Virus-containing supernatants were collected from J2 cells growing in log phase for 24 h in RPMI 1640 supplemented with 20% heat-inactivated fetal calf serum, 12 mM NaHC0 3 , 100 IU/ml penicillin G, 100 g/ml streptomycin, 1 mM sodium pyruvate, 2 mM L-glutamine, 50 M 2-mercaptoethanol, and 12.5 mM HEPES, pH 7.4. Spleens from 12-14-day-old me/me and normal littermate mice were removed and red blood cells were lysed with Tris-NH 4 Cl. Immortalization was carried out by incubating total splenocytes at a concentration of 7 ϫ 10 6 cells/ml with medium containing J2 retrovirus. After 15-30 days, cells that had proliferated were expanded and analyzed for sIgM by flow cytometry. sIgMϩ cells were subsequently cloned by limiting dilution, expanded, and maintained in RPMI 1640 supplemented as described above, except that the fetal calf serum concentration was decreased to 10%.
Flow Cytometry-Cells were stained with the appropriate antibodies in the presence of fully supplemented RPMI. Antibodies used for staining included anti-IgM and anti-Fc␥RIIB (CD32/CD16; 2.4G2), obtained from Pharmingen. Flow cytometric data were obtained on a FACScan using Cell Quest acquisition and analytical software (Becton Dickinson, San Jose, CA).
Calcium Assays-N and ME cells were loaded with 2 M (final concentration) Fura-2 (Molecular Probes, Eugene, OR) in a buffer containing 135 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1 mM MgCl 2 , 5.6 mM glucose, 10 mM HEPES, pH 7.4, and 0.1% bovine serum albumin. Cells were washed once and resuspended at 2 ϫ 10 6 cells/ml. Assays were performed as described previously (21) using a Delta Scan spectrofluorometer (Photon Technologies Inc., South Brunswick, NJ). Cells were activated with 5 g/ml F(abЈ) 2 goat anti-mouse Mu or with 10 g/ml rabbit anti-mouse Mu (Jackson ImmunoResearch) either with or without preincubation with anti-FcR (2.4G2) antibody. To monitor calcium influx, cells were activated with anti-Mu antibodies in the presence of 1.5 mM EGTA, followed by the addition of 5 mM CaCl 2 . The emission values at 510 nm following excitation at 340 or 380 nm were collected and analyzed using FELIX software (Photon Technologies). Analyses of data and generation of graphs were performed using Quattropro.
Cell Activations and Immunoprecipitations-N and ME cells were activated by incubation with 5-10 g/ml F(abЈ) 2 anti-Mu or with 10 g/ml rabbit anti-mouse Mu (Jackson ImmunoResearch) at 37°C. Activations were stopped by addition of lysis buffer containing 1% Nonidet P-40, 150 mM sodium chloride, 5 mM sodium flouride, 1 mM EDTA, 2 mM sodium orthovanadate, 25 mM HEPES, pH 7.4, and the protease inhibitors leupeptin (10 g/ml), aprotinin (10 g/ml), pepstatin (1 g/ml), antipain (1 g/ml), and phenylmethylsulfonyl fluoride (10 g/ml). Typically, 5 ϫ 10 6 cells were activated per time point, and one-tenth of the resulting lysate was removed before immunoprecipitation and analyzed by anti-phosphotyrosine immunoblotting. In some cases, activations were stopped by the addition of ice-cold phosphate-buffered saline, and cell pellets were frozen in liquid N 2 and stored at Ϫ80°C prior to lysis. Lysates were immunoprecipitated for 1-3 h with the indicated antibodies, including anti-CD19 (Pharmingen), anti-Fc␥RIIB (CD32/CD16; Pharmingen), and anti-SHIP (a gift from Dr. Mark Coggeshall, Ohio State University). Anti-Fc␥RIIB antibodies were covalently crosslinked to protein G-agarose beads prior to use. Immunoprecipitates were collected onto either protein A-Sepharose (Pharmacia Biotech Inc.) or protein G-agarose (Oncogene Sciences) beads, washed three times with lysis buffer, separated by SDS-polyacrylamide gel electrophoresis, and transferred onto polyvinylidene difluoride membranes (Immobilon, Millipore).

RESULTS AND DISCUSSION
SHP-1 has been implicated as a negative regulator of antigen receptor (BCR) signaling in B cells and, in particular, in Fc␥RIIB signaling (13,24,25). Comparison of BCR signaling events in primary B cells from me/me mice is difficult because these cells are developmentally distinct and have reduced surface levels of sIgM and Fc␥RIIB compared with age-matched B cells from normal littermates (24). 2 To circumvent such problems, we generated a set of immortalized B cell lines derived from me/me and normal littermate mice, using the murine retrovirus J2, which expresses the viral oncogenes raf and myc (19,20) (Fig. 1A). Previous work indicated that this virus can generate immortal, factor-independent macrophage and B cell lines in vivo and ex vivo (20,26). Unlike B cell lines produced by Abelson MuLV, which are arrested at the pro-or early pre-B cell stage and thus cannot signal through the BCR, many J2-derived lines are sIgMϩ, sIgDϩ, and retain BCR signaling capacity (8,20) (Figs. 2 and 3). We obtained several normal cell lines (N) and one me/me-derived line (ME) (Fig. 1). Southern analysis indicated that these lines are clonal (data not shown), and immunoblotting confirmed that the ME line lacks SHP-1 expression (Fig. 1B). Notably, however, the N and ME lines express equal levels of SHP-2 (Fig. 1B). The surface phenotypes of the N and ME cell lines are similar; of particular importance for the results discussed below, they express comparable levels of sIgM and Fc␥RIIB (Fig. 1, C-F). Both lines also display similar levels of sIgD, CD45 (B220), CD19, and CD22 but do not express CD5, CD11b, B7.1 or B7.2 (data not shown). Therefore, their surface phenotype more closely resembles primary B cells, rather than B-1a, germinal center or memory B cells. Other properties of these cell lines will be discussed in more detail elsewhere. 3 We assessed Fc␥RIIB-mediated inhibition in both cell lines by comparing their calcium response following BCR ligation alone or together with Fc␥RIIB. The change in the ratio (chelated/free) of Fura-2 fluorescence was used to measure calcium responses to both stimuli. We first tested the N cell line to determine whether J2 immortalization interferes with BCR activation and/or Fc␥RIIB-mediated inhibition. N cells activated with F(abЈ) 2 anti-Mu (engagement of BCR alone) exhibited a robust increase in intracellular free calcium. This response was markedly diminished in N cells stimulated with intact anti-Mu ( Fig. 2A). Preincubation with anti-Fc␥RIIB antibodies (2.4G2) abrogated the ability of intact anti-Mu to inhibit the calcium response in N cells (Fig. 2B), indicating that inhibition occurred solely via engagement of Fc␥RIIB. N cells activated in the presence of EGTA showed little difference in peak calcium release from intracellular stores (Fig. 2C). Readdition of CaCl 2 after the release phase to allow measurement of calcium influx revealed strong inhibition by intact anti-Mu (Fig. 2C). These results, which are consistent with those reported for primary B cells and several B cell lines (see Introduction), demonstrate that when engaged, Fc␥RIIB inhibits BCR signaling in the N cell line by preventing calcium influx. Importantly, the cross-linking antibodies, although derived from different species, evoked identical peak levels of calcium release, indicating that they provide comparable activation stimuli (Fig. 2C). Likewise, the calcium responses to F(abЈ) 2 and intact anti-Mu in the presence of anti-Fc␥RIIB (2.4G2) antibodies were virtually identical (Fig. 2, compare A with B). Recently, we have confirmed these results using F(abЈ) 2 and intact anti-Mu IgG from the same species (rabbit) as crosslinking reagents (data not shown). Taken together, these data show that J2-immortalized B cells provide a valid model to study BCR signal transduction and its regulation by Fc␥RIIB.
To determine whether SHP-1 is required for negative regulation of calcium influx by Fc␥RIIB, we compared calcium responses in ME cells following ligation of the BCR with either F(abЈ) 2 or intact anti-Mu. Surprisingly, ME cells activated with intact anti-Mu also showed complete inhibition of the calcium response compared with those stimulated with F(abЈ) 2 anti-Mu (Fig. 2D); again, inhibition was reversed by preincubation with anti-Fc␥RIIB antibodies (Fig. 2E). As in N cells, there was little difference in peak calcium release from intracellular stores following ligation of ME cells with either antibody in the presence of EGTA, but extracellular calcium influx was markedly reduced with intact anti-Mu compared with F(abЈ) 2 anti-Mu (Fig. 2F). As with N cells, we have recently confirmed these results using matched rabbit F(abЈ) 2 and intact anti-Mu IgG (data not shown). Because ME cells exhibit Fc␥RIIB-mediated inhibition of the BCR-induced calcium response, we conclude that, as in mast cells (17), SHP-1 is not required for FcRmediated inhibition of B cell signaling.
Although SHP-1 is dispensable for Fc␥RIIB-mediated inhibition of BCR calcium influx, it remained possible that it influences Fc␥RIIB-evoked tyrosyl phosphorylation events. To address this issue, we first examined the tyrosyl phosphorylation of Fc␥RIIB itself. Previous studies established that tyrosyl phosphorylation of Fc␥RIIB occurs upon co-cross-linking to the BCR, but not when the BCR alone is engaged (12). Anti-phosphotyrosine immunoblots of Fc␥RIIB immunoprecipitates from N and ME cells activated by either F(abЈ) 2 or by intact anti-Mu ligation revealed Fc␥RIIB tyrosyl phosphorylation only upon co-cross-linking (Fig. 3A, upper panel). Fc␥RIIB was not signif-icantly tyrosyl phosphorylated either basally or upon F(abЈ) 2 anti-Mu stimulation in ME cells, indicating that loss of SHP-1 does not lead to aberrant Fc␥RIIB tyrosyl phosphorylation (Fig.  3A, upper panel). To determine if absence of SHP-1 resulted in global differences in total cellular tyrosyl phosphorylation upon Fc␥RIIB/BCR co-cross-linking, we compared anti-phosphotyrosine immunoblots of total cell lysates from cells that were activated with either F(abЈ) 2 or intact anti-Mu. In N cells, each stimulus yielded a comparable overall total cellular tyrosyl phosphorylation response (Fig. 3C), consistent with previous studies that reported no gross difference in total tyrosyl phosphorylation between intact versus F(abЈ) 2 anti-BCR stimulation (12). Each stimulus also evoked comparable increases in total cellular phosphotyrosine in ME cells (Fig. 3C). A detailed comparison of BCR signal transduction between N, ME, and ME cells reconstituted with wild type and mutant SHP-1 will be presented elsewhere. 3 CD19 has been reported to display decreased tyrosyl phosphorylation upon Fc␥RIIB co-cross-linking with the BCR. It has been suggested that SHP-1 mediates the differential tyrosyl phosphorylation of this co-receptor (15). To address these issues, we examined CD19 tyrosyl phosphorylation in both cell lines. In N cells, anti-phosphotyrosine immunoblotting of CD19 immunoprecipitates revealed that tyrosyl phosphorylation of CD19 is slightly decreased when the BCR is co-cross-linked with Fc␥RIIB (Fig. 3D), consistent with the previous study (15). However, decreased CD19 tyrosyl phosphorylation also was observed when the BCR was co-cross-linked with Fc␥RIIB in ME cells (Fig. 3D). These data suggest that SHP-1 is not solely responsible for CD19 dephosphorylation in response to Fc␥RIIB engagement.
Our results clearly establish that SHP-1 is not required for Fc␥RIIB-mediated inhibition of BCR-induced calcium influx. These findings contrast with an earlier report, which suggested that SHP-1 is necessary for Fc␥RIIB-mediated inhibition in B cells (13). The earlier study compared signaling in primary lymphocytes from normal and me/me or me v /me v mice. Primary B cells from me/me mice are developmentally distinct from those of their normal counterparts and are similar in phenotype to anergized B cells, exhibiting decreased levels of BCR and increased expression of activation markers (24). The failure of the earlier workers to observe effects of Fc␥RIIB engagement on the proliferative responses of me/me B cells may reflect developmental differences caused by the lack of SHP-1, rather than the direct effect of absence of SHP-1 on signaling via Fc␥RIIB. Alternatively, the high levels of circulating immune complexes in me/me mice may lead to desensitization of Fc␥RIIB signaling (although me/me B cells do retain surface Fc␥RIIB; data not shown).
Our results also establish that SHP-1 is not required for the Fc␥RIIB-evoked decrease in CD19 tyrosyl phosphorylation. Presumably, because there has been no demonstrated change in any protein-tyrosine kinase activity upon Fc␥RIIB engagement (12), a PTP other than SHP-1 promotes CD19 dephosphorylation. SHP-2 has been shown to bind the Fc␥RIIB ITIM in vitro (14), raising the possibility that it may be responsible for CD19 dephosphorylation in vivo. In any case, the functional significance of CD19 dephosphorylation in response to Fc␥RIIB engagement remains unclear.
As in mast cells (17) and other B cell lines and primary B cells, SHIP, an SH2-domain containing inositol phosphatase, associates with tyrosyl phosphorylated Fc␥RIIB and becomes tyrosyl phosphorylated in both N and ME cells (Figs. 3, A, lower  panel, and B), raising the possibility that it mediates inhibitory signaling by this receptor. The mechanism by which an inositol phosphatase could promote inhibition of calcium influx is unclear. Hydrolysis of phosphatidylinositol 3,4,5-trisphosphate and/or phosphatidylinositol 4,5-bisphosphate could directly or indirectly modulate a calcium channel responsible for influx. Such a channel(s) remains to be identified in B cells. Dephosphorylation of CD19, with consequent diminished recruitment of PI-3K (15), might lead to similar effects on phosphatidylinositol-containing lipids, as recruitment of SHIP to Fc␥RIIB. Therefore, it remains possible that Fc␥RIIB may function by FIG. 3. Absence of SHP-1 has no effect on biochemical events evoked by Fc␥RIIB engagement. A, anti-phosphotyrosine immunoblot of anti-Fc␥RIIB immunoprecipitates from N and ME cell lines stimulated with either F(abЈ) 2 anti-Mu (5 g/ml) or intact anti-Mu (10 g/ ml). This immunoblot also was stripped and reprobed with anti-SHIP antibodies (arrow). B, anti-phosphotyrosine immunoblot of anti-SHIP immunoprecipitates from N and ME cells stimulated with either F(abЈ) 2 anti-Mu (5 g/ml) or intact anti-Mu (10 g/ml). Note co-immunoprecipitation of tyrosyl phosphorylated Fc␥RIIB. C, anti-phosphotyrosine immunoblot of total cell lysates (40 g) from N and ME cells, stimulated as indicated. Note that the tyrosyl phosphorylation of an approximate 50-kDa protein with intact anti-Mu is Fc␥RIIB. As a control for protein loading, the blot was reprobed with anti-SHP-2 antibodies (lower panel). D, anti-phosphotyrosine immunoblot of CD19 immunoprecipitates from N and ME cells activated with either F(abЈ) 2 anti-Mu (5 g/ml) or intact anti-Mu (10 g/ ml) for the indicated times. A duplicate aliquot of each CD19 immunoprecipitate was subjected to anti-CD19 immunoblotting to show that comparable amounts of CD19 were being analyzed (reprobing was not possible using this antibody). Note that CD19 tyrosyl phosphorylation decreases upon Fc␥RIIB engagement in both cell lines. Unst., unstimulated. recruiting both SHIP and at least SHP-1 or SHP-2. However, it is important to note that we have been unable to detect stable association of SHP-2 with tyrosyl phosphorylated Fc␥RIIB in B cells (data not shown). Clarifying the roles of SHIP and SHP-2 in Fc␥RIIB signaling in B cells will most likely require generation of cell lines or mice that lack expression of these proteins.