Co-aggregation of FcγRII with FcϵRI on Human Mast Cells Inhibits Antigen-induced Secretion and Involves SHIP-Grb2-Dok Complexes*

Signaling through the high affinity IgE receptor FcϵRI on human basophils and rodent mast cells is decreased by co-aggregating these receptors to the low affinity IgG receptor FcγRII. We used a recently described fusion protein, GE2, which is composed of key portions of the human γ1 and the human ϵ heavy chains, to dissect the mechanisms that lead to human mast cell and basophil inhibition through co-aggregation of FcγRII and FcϵRI. Unstimulated human mast cells derived from umbilical cord blood express the immunoreceptor tyrosine-based inhibitory motif-containing receptor FcγRII but not FcγRI or FcγRIII. Interaction of the mast cells with GE2 alone did not cause degranulation. Co-aggregating FcϵRI and FcγRII with GE2 1) significantly inhibited IgE-mediated histamine release, cytokine production, and Ca2+ mobilization, 2) reduced the antigen-induced morphological changes associated with mast cell degranulation, 3) reduced the tyrosine phosphorylation of several cellular substrates, and 4) increased the tyrosine phosphorylation of the adapter protein downstream of kinase 1 (p62dok; Dok), growth factor receptor-bound protein 2 (Grb2), and SH2 domain containing inositol 5-phosphatase (SHIP). Tyrosine phosphorylation of Dok was associated with increased binding to Grb2. Surprisingly, in non-stimulated cells, there were complexes of phosphorylated SHIP-Grb2-Dok that were lost upon IgE receptor activation but retained under conditions of Fcϵ-Fcγ co-aggregation. Finally, studies using mast cells from Dok-1 knock-out mice showed that IgE alone triggers degranulation supporting an inhibitory role for Dok degranulation. Our results demonstrate how human FcϵRI-mediated responses can be inhibited by co-aggregation with FcγRIIB and implicate Dok, SHIP, and Grb2 as key intermediates in regulating antigen-induced mediator release.

Mast cells generate pro-inflammatory mediators that initiate and propagate allergic inflammation. IgE-mediated mediator release is initiated through the high affinity IgE receptor Fc⑀RI, 1 in which the ␣ subunit has IgE binding activity and the ␤ and ␥ subunits have signaling activity. Signaling is primarily mediated by immunoreceptor tyrosine-based activation motifs (ITAMs) located in the carboxyl-terminal cytoplasmic tails of each of the ␤ and ␥ subunits. ITAMs are typically 26-or 27amino acid stretches including two YXXL motifs separated by 9 or 10 amino acids. Cross-linking IgE-primed receptors with a multivalent antigen results in the activation of Lyn, which in turn phosphorylates ITAMs, creating docking sites for Syk kinase family members and permitting signal propagation (for review, see Ref. 1).
Many immune cells that express ITAM-containing antigen receptors also express receptors with a related immunoreceptor tyrosine-based inhibitory motifs (ITIMs). ITIMs are 13-amino acid sequences containing a single YXXL motif (2). They were first recognized in the cytoplasmic tails of the Fc␥RIIB isoform (Fc␥RIIB1) of rodent B cells. Their negative signaling activity was revealed by studies showing that co-aggregating the BCR with Fc␥RIIB1 inhibits the BCR-mediated activation of rodent B cells (3). Studies in murine bone marrow derived mast cells (BMMCs) showed that co-aggregating murine Fc␥RIIB to Fc⑀RI also inhibits IgE-induced mast cell degranulation and that this inhibition is dependent on the intact intracytoplasmic tail of the Fc␥RIIB (4). Further studies using RBL-2H3 cells transfected with cDNA encoding human wild type and mutant Fc␥RIIB showed that the intracytoplasmic domain responsible for the negative signaling is the ITIM found in Fc␥RIIB (5). Subsequently, it was shown that the phosphorylation of the mast cell Fc␥RIIB-ITIM is mediated by Lyn and results in the recruitment of specific tyrosine phosphatases (6).
The negative regulation of signaling through ITAM-containing receptors by signals generated by the co-cross-linked ITIM-containing Fc␥RIIB isoforms has been demonstrated in mouse and human B cells that naturally co-express the BCR and Fc␥RIIB1, in T cell and rodent mast cell lines that express the T cell receptor or Fc⑀RI plus endogenous or transfected Fc␥RIIB1 (5), and in human basophils (7). Negative regulation by other ITIM-containing receptors has been demonstrated in several other model systems (8). It is not clear whether human mast cells express ITIM receptors and whether IgE-mediated responses can be down-regulated through their activity. Although there has been a great effort to discover and study ITIM receptors, little progress has been made in harnessing their inherent inhibitory properties for potential therapeutic interventions.
We recently showed that a novel human Fc⑀-Fc␥ receptorbinding fusion protein (GE2) could inhibit human basophil functional and biochemical responses in vitro and in vivo (9,64). Subsequently, it has been demonstrated that GE2 can inhibit IgE production by human B cell (10) and cytokine production from Fc⑀RI/Fc␥RII-expressing, Langerhan cell-like dendritic cells (11).
Here we used umbilical cord blood-derived human mast cells (CBMCs) to test the mechanisms of inhibition by Fc⑀RI-Fc␥RII co-aggregation on human mast cells. We demonstrated that human mast cells express Fc␥RII (CD32) but not Fc␥RI (CD64) or Fc␥RIII (CD16). We provide the first evidence that the functional responses of human mast cell Fc⑀RI can be downregulated by Fc␥RII. The inhibition involves alterations in protein tyrosine phosphorylation, including Syk. We provide evidence that Grb2, Dok, and SHIP phosphorylation is associated with mast cell inactivation and Fc␥RII-mediated inhibition of Fc⑀RI responses. Although previous rodent studies have implicated SHIP, our findings further implicate Grb2 and Dok as additional "gatekeepers" of human Fc⑀RI signaling.
Umbilical CBMCs-Umbilical cord blood was collected from normal, full-term deliveries as approved by the Human Studies Committees at the Medical College of Virginia. EDTA-treated (0.01% final) cord blood was layered over Ficoll-Hypaque (density 1.077; Sigma) 1:1 (v/v), and the tubes were centrifuged at 350 ϫ g for 20 min at room temperature with no brake. Mononuclear cells recovered at the interphase were washed with phosphate-buffered saline with 0.1% bovine serum albumin (PBS/BSA). If red blood cell contamination was present, the pellet was resuspended in PBS/BSA and was relayered over the Ficoll (1:1, v/v), centrifuged, and washed as above. The CD34ϩ stem cells were isolated using a negative selection mixture and magnetic separation of non-CD34-positive cells using the technique given by the manufacturer (Miltenyi, Auburn, CA). The cells were suspended at a concentration of ϳ0.5-1.0 ϫ 10 6 cells/ml in Iscove's modified Dulbecco's medium (Sigma) supplemented with 10% fetal calf serum (Hyclone, Logan, UT), 100 ng/ml recombinant human stem cell factor and 50 ng/ml recombinant human interleukin-6 (IL-6) (both from BIOSOURCE, Camarillo, CA), 10 M Hepes, 50 M 2-mercaptoethanol, 4 mM L-glutamine, 1ϫ minimum Eagle's medium amino acids, 1ϫ vitamin solution (Invitrogen), 1 mM sodium pyruvate, 100 units/ml penicillin, and 100 g/ml streptomycin (13). To enhance Fc⑀RI expression, human anti-NIP IgE (1 g/ ml) was added at least 1 week prior to each experiment (13). Cultures were maintained at 5% CO 2 and 37°C, and the medium was changed every 4 -5 days with the cells placed back into the original flask. Cells were used at 6 -9 weeks. As reported previously using a similar protocol (13), the purity of the mast cells was always greater than 90% as determined by tryptase immunostaining (14) (data not shown). Aliquots of cells were removed on different days and before use for the analysis of viability (by trypan blue exclusion) and for cytospin preparation.
Flow Cytometry-Cells were recovered by centrifugation at 800 ϫ g at 4°C, washed with PBS/BSA, and incubated for 30 min at 4°C with a 1:500 dilution of normal human serum. The cells were washed and incubated with the indicated Abs (10 g/ml) for 1 h at 4°C. After antibody labeling, the cells were washed and incubated with a 1:100 dilution of F(abЈ) 2 -fluorescein isothiocyanate-goat anti-mouse for 30 min at 4°C. After three washes, cells were resuspended in 400 l of PBS/BSA. The mean intensity fluorescence was determined for at least 10,000 cells using a Coulter Epics Elite flow cytometer. MOPC31C nonspecific mouse IgG or non-immune rabbit IgG was used as a negative control. All experiments were performed in duplicate.
Cell Activation-CBMCs were suspended in fresh medium (without cytokines) and sensitized with 10 g/ml human anti-NIP IgE overnight at 37°C in a 5% CO 2 incubator. The next morning, GE2 was added for at least 2 h to a final concentration indicated in the "Results" section. Cells were washed and activated, and mediator release was measured as described previously (15). Transmission Electron Microscopy-Suspensions of IgE-sensitized CBMCs were challenged with or without GE2 (10 g/ml) as above, washed, and incubated at 37°C in prewarmed HBSS ϩ (HBSS with 1 mM CaCl 2 and 1 mM MgCl 2 ) without or with NIP/BSA (200 ng/ml) for 30 min. Cells were collected by centrifugation, and pellets were fixed in 2% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4, for 30 min at room temperature, rinsed with cacodylate, and processed as described previously (16). Thin sections were observed using a Hitachi 600 transmission electron microscope.
Lysis, Immunoprecipitation, and Immunoblotting-IgE-sensitized mast cells, with or without GE2, were activated as above. Preparation of cell lysates, immunoprecipitation, and Western blotting were performed as described previously (7).

Identification of Fc␥Rs on
CBMCs-ITIM-containing receptors are an expanding family of immune inhibitory receptors that have the ability to inhibit immune responses (8). We are not aware of studies examining their expression on human mast cells. As seen in Fig. 1, CBMCs express, in addition to Fc␥RIIB, the ITIM-containing receptor CD22 and the ITIMlike-containing receptor CD81. Similar to what we have shown previously for human basophil Fc␥Rs, human mast cells exclusively express the low affinity IgG receptor Fc␥RII. Thus CBMCs react strongly with the pan-anti-Fc␥RII mAb AT10 (Fig. 1). Confirming earlier studies examining cultured human mast cells, little or no Fc␥RI or Fc␥RIII expression was detected (17). Thus, cultured human mast cells express only the low affinity IgG receptor Fc␥RII. There are multiple Fc␥RII isoforms, representing the products of three distinct genes (18). Results of reverse transcription-PCR analysis using previously described methods (7,19) showed that CBMCs contain transcripts for the ITIMcontaining Fc␥RII isoform Fc␥RIIB and the ITAM-containing Fc␥RII isoforms Fc␥RIIA and Fc␥RIIC (data not shown). However, we did not detect surface expression of Fc␥RIIA by FACS analysis using the Fc␥RIIA-specific Ab IV.3 (data not shown). In addition, no Fc␥RIIA was detected by Western blot analysis in mast cell lysates using an Ab raised against the cytoplasmic tail of Fc␥RIIA (Ab260), whereas strong signals were detected when using an Fc␥RIIB-specific Ab (Ab163; a gift of Dr. Susheela Tridandapani, Ohio State University, Columbus, OH) ( Fig.  1, inset) (20). Thus, although CBMCs express mRNA for Fc␥RIIA, it appears that only Fc␥RIIB is expressed on the cellular membrane in unstimulated, human CBMCs.

Fc␥RII-Fc⑀RI Co-aggregation Inhibits Fc␥RI-dependent Human Mast Cell Degranulation-Previous investigators have established that Fc␥RII-Fc⑀RI co-aggregation can inhibit
Fc⑀RI responses in a variety of model systems. We designed and described an Fc␥R-Fc⑀R-binding chimeric protein that can directly induce this interaction (GE2; Ref. 9). Fig. 2A shows the secretory responses induced by adding optimal concentrations of NIP/BSA (200 ng/ml) to CBMCs that had been sensitized with human anti-NIP IgE and treated with the GE2 fusion protein. GE2 incubation for 2 h prior to antigen stimulation inhibited IgE-mediated histamine release in a dose-dependent fashion. This inhibition could not be accounted for by competition between IgE and GE2 for Fc⑀RI binding as substitution of nonspecific purified IgE myeloma protein (PS-IgE) for the GE2 resulted in IgE-mediated release comparable with non-GE2challenged control cells. Similar to what we reported previously using human basophils (9), no secretion is induced by adding chimeric protein alone to IgE-sensitized and unsensitized cells (data not shown). In four separate experiments, co-aggregating Fc⑀RI and Fc␥RII with GE2 (10 g/ml) reduced Fc⑀RI-mediated histamine secretion by an average of 68%.
Fc␥RII-Fc⑀RI Co-aggregation Inhibits CBMCs Fc⑀RI-medi-ated TNF-␣ Production-Human mast cells produce TNF-␣ following IgE-mediated stimulation (21). The results in Fig. 2B show that anti-NIP IgE-sensitized CBMCs that were treated with GE2 released significantly less TNF-␣ in response to 12-h stimulation with NIP/BSA compared with sensitized but GE2untreated cells. The average inhibition of Fc⑀RI-mediated TNF-␣ production by Fc␥RII-Fc⑀RI co-aggregation was 60 and 64% using 5 and 10 g/ml, respectively.

Fc␥RII-Fc⑀RI Co-aggregation Inhibits CBMCs Fc⑀RI-mediated Ca 2ϩ
Mobilization-Previous studies have indicated that human mast cell Fc⑀RI-mediated degranulation and cytokine production both depend on the release of Ca 2ϩ from intracellular stores and on the subsequent sustained influx of extracellular Ca 2ϩ (22,23). We used the Ca 2ϩ indicator dye Indo-1 and FACS analysis to examine the pattern of Fc⑀RI-mediated Ca 2ϩ mobilization in human mast cells following GE2 treatment (Fig. 3).
Cross-linking anti-NIP IgE-primed Fc⑀RI with NIP/BSA leads, after a lag phase, to a rapid increase in intracellular Ca 2ϩ levels (Fig. 3A). This Ca 2ϩ spike is followed by a sustained elevation in Ca 2ϩ levels, the Ca 2ϩ plateau, which persists for up to 3 min. In the experiment shown in Fig. 3A, the lag time from antigen addition to Fc⑀RI-mediated Ca 2ϩ influx was ϳ24 s. When cells were treated with GE2 prior to antigen challenge, the lag time from antigen addition to the initial rise in [Ca 2ϩ ] more than doubled to 60 s. Thus, co-aggregating Fc␥RII and Fc⑀RI with GE2 decreases the initial Ca 2ϩ mobilization in human mast cells.
To further investigate whether co-aggregation inhibits the release of intracellular calcium stores, we utilized thapsigargin, which induces an increase in [Ca 2ϩ ] i because of store leakage from the endoplasmic reticulum. As seen in Fig. 3B, thapsigargin induced a sustained increase in [Ca 2ϩ ] i . However, when GE2 (10 g/ml) and antigen were added prior to challenge with thapsigargin, the increase observed in [Ca 2ϩ ] i was smaller, and a sustained phase of Ca 2ϩ was lacking. These findings indicate that the effect of GE2 appeared to be mainly due to an inhibition of the endoplasmic reticulum Ca 2ϩ store as well as sustained, store-dependent [Ca 2ϩ ] i influx.
Fc␥RII-Fc⑀RI Co-aggregation Inhibits CBMCs Fc⑀RI-mediated Morphological Changes-Under transmission electron microscopy, unstimulated cultured human mast cells (Fig. 4A) show large unsegmented nuclei and relatively sparse endoplasmic reticulum and mitochondria. Granules are numerous and contain loosely packed matrix material plus occasional dense core material, multilamellar membranes, and internal vesicles. Fig. 4B shows the intracellular fusion of granules (arrows) and the extracellular release of granule contents associated with normal IgE-mediated mast cell activation (24 -26). In contrast, the granules in cells treated with GE2 remain individual and intact following antigen stimulation (Fig. 4C).
Fc␥RII-Fc⑀RI Co-aggregation Affects Fc⑀RI-mediated Tyrosine Phosphorylation of Multiple Substrates-Protein tyrosine phosphorylation is a key event linking Fc⑀RI cross-linking to downstream signaling in human mast cells and basophils (15,16,23). To investigate the effects of co-aggregating Fc␥RII-Fc⑀RI on IgE-mediated protein tyrosine phosphorylation, whole cell lysates of IgE-activated cells, with or without GE2 preincubation, were resolved by SDS-PAGE and probed with anti-Tyr(P) antibodies. As shown in Fig. 5A, NIP/BSA crosslinking of anti-NIP IgE-sensitized Fc⑀RI induced the tyrosine phosphorylation of several different substrates at 2, 5, and 15 min (lanes 2, 4, and 6) compared with non-activated cells (lane 1). The most prominent signals were species appearing at ϳ110, 85, 76, 72, 66, 44, 42, 38, and 28 kDa. When cells were treated with GE2, tyrosine phosphorylation of several of these substrates was reduced. These included proteins with molecular masses of ϳ85, 72, 66, 44, and 38 kDa (indicated with arrows). Thus, the down-regulation of IgE-mediated degranulation by co-aggregation of Fc␥RII and Fc⑀RI by GE2 appears to involve the reduction of tyrosine phosphorylation of several different proteins.
Previous investigations have shown that the tyrosine kinase Syk is quickly phosphorylated in IgE-stimulated human Fc⑀RIpositive cells (15,23,27). In human basophils, co-aggregation of Fc⑀RI and Fc␥RII with antigen-specific IgE and IgG (7) or GE2 (9) inhibits the tyrosine phosphorylation of Syk. As seen in   5B using whole cell lysates of CBMCs, co-aggregation by GE2 reduced the tyrosine phosphorylation of a protein with the approximate molecular mass of Syk (ϳ72 kDa). Thus, we wanted to examine specifically the phosphorylation of Syk in GE2-treated CBMCs after IgE stimulation. Cells with or without GE2 preincubation were stimulated for the indicated periods and lysed, and anti-Syk immunoprecipitates were analyzed by immunoblotting with anti-Tyr(P) Ab. As seen in Fig. 5B Syk was tyrosine-phosphorylated within 2 min after addition of antigen. The amount of Syk tyrosine phosphorylation was markedly reduced in IgE-activated cells that had been treated with GE2 compared with non-treated IgE-activated cells.
Grb2 Phosphorylation Is Increased in Resting and Fc␥RII-Fc⑀RI-co-aggregated Cells-Grb2 is an adapter protein associated with Fc⑀RI and Fc␥RIIB signaling (28,29). Initial experiments indicated Grb2 was differentially phosphorylated and associated with other signaling molecules under Fc⑀RI-Fc␥RIIB-co-aggregating conditions (not shown). Grb2 was immunoprecipitated from lysates of variously activated cells, and immune complexes were resolved by SDS-PAGE and probed FIG. 3. A, Fc␥RII-Fc⑀RI co-aggregation inhibits Fc⑀RI-mediated Ca 2ϩ mobilization in CBMCs. Cells were sensitized with 10 g/ml of anti-NIP IgE with or without GE2, loaded with Indo-1, and challenged with 200 ng/ml NIP/BSA for 4 min. Change in absorbance was monitored using FACS analysis as described under "Experimental Procedures." Each plot represents the Ca 2ϩ response of whole cell populations. The data are representative of two separate experiments. B, GE2 inhibits the release of intracellular calcium stores. Fura-2/AM-loaded cells were challenged with thapsigargin (thaps) with or without preincubation with GE2 (10 g/ml). Change in absorbance was monitored as described under "Experimental Procedures." XL, cross-linker.
with anti-Tyr(P) Abs. As seen in Fig. 6, in unstimulated cells Grb2 immunoprecipitated tyrosine-phosphorylated proteins with apparent molecular masses of 60 -65, 54 -60, 35-37, and 20 -25 kDa. Upon antigen-driven Fc⑀RI cross-linking, the interaction of Grb2 with the phosphorylated species at 20 -25 and 60 -65 kDa was significantly reduced. However, these reductions were not observed when antigen-treated cells had been treated with GE2 so as to drive Fc⑀-Fc␥ co-aggregation. Stripping and reprobing with anti-Grb2 antibodies determined that the molecular mass species at about 25 kDa was Grb2 (Fig. 6,  bottom blot). The increased phosphorylation of Grb2 in resting versus IgE-stimulated cells suggests that Grb2 phosphorylation is important for preventing Fc⑀RI activation and for inhibitory signaling.
Phosphorylated Grb2 and Dok Complexes Are Increased in Resting and Fc␥RII-Fc⑀RI-co-aggregated Cells-We hypothesized that the 60 -65-kDa species complexed to Grb2 was Dok. Grb2 was immunoprecipitated, and the blots were probed with anti-Tyr(P) Abs. Confirming the results seen in Fig. 6, top blot, the phosphorylation of the 60 -65-kDa species was reduced under IgE-cross-linking conditions. However, in resting and co-aggregated cells a stronger phosphorylation signal was observed (Fig. 7, top). Probing with anti-Dok Abs revealed that the Dok-Grb2 interaction was incrementally lost upon IgE activation but was maintained and slightly increased in resting and Fc⑀-Fc␥-co-aggregated cells (Fig. 7, middle). This difference was not because of unequal protein loading as shown in the Grb2 blot (Fig. 7, bottom). Studies are underway to deter-mine the identity of the 35-40-and 50 -55-kDa proteins, but linker for activation of T cells (LAT) and Shc are likely candidates, respectively. Nonetheless, Grb2 and Dok appear to form a phosphorylation-dependent complex in resting mast cells that is lost upon activation but is maintained and slightly increased under Fc⑀-Fc␥ inhibitory signaling. (30). Dok was immunoprecipitated from cells as described above, and blots were probed with anti-Tyr(P) Abs. As seen in Fig. 8A, a phosphorylated protein shown to be SHIP (middle) at an apparent molecular mass of 145 kDa was differentially tyrosine-phosphorylated in resting, Fc⑀-, and Fc⑀-Fc␥-stimulated cells (top). Specifically, the strong basal tyrosine phosphorylation was decreased in IgE-stimulated cells at 2, 5, and 15 min. However, the tyrosine phosphorylation was increased in Fc⑀-Fc␥-stimulated cells at 5 and 15 min. In addition, immunoblot analysis of Dok immunoprecipitations using antiphosphotyrosine Abs demonstrated that Dok is constitutively phosphorylated on tyrosine at a low level in unstimulated cells, and phosphorylation is enhanced following aggregation of Fc⑀RI alone. Ras-GAP was found to associate with Dok in unstimulated and GE2-co-aggregated cells (Fig. 8B). Thus, in human mast cells, tyrosine-phosphorylated SHIP may exert its gatekeeper role (31) through its interaction with Dok. By binding to Dok, SHIP may negatively regulate Ras function by enhancing its intrinsic GTPase activity through RasGAP.

Phosphorylated SHIP and Dok Complexes Are Increased in Resting and Fc␥RII-Fc⑀RI-co-aggregated Cells-Rodent studies suggest that SHIP may mediate Fc␥RII inhibitory function in mast cells through the recruitment of Dok
IgE Sensitization Alone Induces Degranulation in Dok Knock-out Bone Marrow Mast Cells-The data presented above suggest that Dok may play a role in setting the threshold for degranulation in mast cells. To determine whether Dok was involved in maintaining Fc⑀RI in a quiescent state, we investigated the effects that IgE binding alone had on BMMCs from Dok knock-out mice that have been described previously (32,33). As seen in Fig. 9, degranulation of Dok knock-out BMMCs was observed simply with the addition of IgE clones from either of two different murine clones. Degranulation from these cells did not follow exposure to mouse IgG, nor did wild type BMMCs undergo degranulation following exposure to the same IgE. Taken together, these studies demonstrate a vital role for Dok in setting the threshold for degranulation in mast cells. DISCUSSION The ability of Fc␥RII to down-regulate human basophil Fc⑀RI-induced degranulation was first reported by Daeron and co-workers (4) and confirmed in later studies (7). Based on our recent studies using human basophils and a novel chimeric human Fc␥-Fc⑀ Ig protein (9), we chose to focus on the use of ITIM/ITAM-receptor cross-linking as a way to block the initiation of the Fc⑀RI response in mast cells. Because mast cells are clearly a distinct lineage from basophils in humans and have a number of distinctive functional properties in mediating human disease (34), it is very important to demonstrate their specific response to Fc␥RII-Fc⑀RI co-aggregation. We report here that co-aggregating Fc␥RII and Fc⑀RI on human mast cells inhibits Fc⑀RI-induced mediator release. These results provide the first demonstration of antigen-specific down-regu- lation of Fc⑀RI-mediated signaling though Fc␥RII-Fc⑀RI coaggregation on human mast cells.
Here, we confirm that human mast cells express CD32 (Fc␥RII) but not CD16 (Fc␥RIII) or CD64 (Fc␥RI). Although CBMCs contained RNA transcripts for all three Fc␥RII isoforms, we were not able to detect Fc␥RIIA protein expression by FACS analysis with Ab IV.3 or by Western blotting with an Ab raised against the cytoplasmic tail of Fc␥RIIA. It is still possible that there is donor variability in Fc␥RII expression in human mast cells as has been shown in human monocytes (35), platelets (36), and NK cells (19). Human monocytes can also alter Fc␥RII isoform expression in response to certain cyto-kines and culture conditions (37). Indeed, we have encountered periodically human Fc⑀RI-positive cell preparations that did not respond to negative signaling, and we are investigating the relationship between Fc␥RII isoform expression and the ability of CBMCs to undergo co-aggregation and inhibition of Fc⑀RI signaling.
Granule secretion and cytokine production are both Ca 2ϩ -dependent processes in human mast cells. We found the lag time between antigen addition and initial Ca 2ϩ mobilization very similar to those reported previously using cultured human mast cells (23,38). However, as with human basophils (7), GE2-driven Fc␥RII-Fc⑀RI co-aggregation extended the lag time from Fc⑀RI cross-linking to the Ca 2ϩ spike response and reduced the magnitude of the Ca 2ϩ plateau. In addition, GE2 appears to inhibit the release of intracellular calcium stores as thapsigargin-induced release was inhibited by GE2. Co-aggregation of Fc␥RII-BCR on B cells under a variety of co-crosslinking conditions results in premature termination of Ca 2ϩ influx, sometimes associated with a reduced Ca 2ϩ spike response but not an increased delay to the spike response (39 -42). This difference between mast cells/basophils and B cells is unexplained but likely reflects distinct signaling properties of B cells and human Fc⑀RI-positive cells and particularly the presence of Fc␥RIIA on resting and activated B cells.
Tyrosine phosphorylation of cellular substrates is an integral part of IgE-mediated signaling in human mast cells (38) and basophils (7). Several substrates were differentially phospho-rylated upon antigen activation in CBMCs (Fig. 5A). We reported previously that the tyrosine kinase Syk is critical in human basophil IgE-mediated signal transduction (15,16) and that co-aggregation inhibits Syk phosphorylation (7,9). Here we show little or no tyrosine-phosphorylated Syk present in resting mast cells, whereas Fc⑀RI cross-linking induced a large increase in Syk phosphorylation (Fig. 5B). The kinetics of phosphorylation observed upon IgE activation were similar to previous reports using human mast cells (23). As with human basophils, Syk phosphorylation was markedly decreased by co-aggregating Fc␥RII and Fc⑀RI.
Surprisingly, we observed increased Grb2 phosphorylation in resting and Fc␥RII-Fc⑀RI-co-aggregated cells with reduced Grb2 phosphorylation seen in mast cells activated through Fc⑀RI alone. We are not aware of any studies that associate Grb2 tyrosine phosphorylation with inhibitory Fc⑀RI signaling. Grb2 functions as an adaptor protein associated with proteintyrosine kinase signaling (43). Published studies, mostly nonphysiological, focus on the phosphorylation of other signaling intermediates, which then leads to the binding of Grb2 SH2 and SH3 domains. The tyrosine phosphorylation of Grb2 itself and the effect on signaling responses have not been characterized, especially in human mast cells. However, Grb2 protein has several tyrosines capable of being phosphorylated (44), and Grb2 tyrosine phosphorylation has been detected under physiological conditions in several cell types (45)(46)(47). Analogous to our findings, Grb2 tyrosine phosphorylation was clearly shown to negatively regulate the downstream activation of tyrosine kinase signaling pathways in other cell types (48). Our data support these previous studies and suggest that tyrosine phosphorylation of Grb2 may be an important step in a novel mechanism of down-regulation of Fc⑀RI activation of human mast cells.
We also found that Dok is involved in IgE receptor signaling in human mast cells. Similar to Grb2, Dok phosphorylation was reduced in Fc⑀RI-stimulated cells, whereas it was maintained in co-aggregated "inhibited" cells. The increase in Dok tyrosine phosphorylation was also associated with increased interaction with Grb2. Previous studies have shown that Grb2 SH2 domains can bind to phosphorylated Dok (and a phosphorylated 30 -40-kDa protein, as we also found) (Fig. 6) (49). Reports with transfected RBL-2H3 cells demonstrated that co-aggregation of Fc␥RII with Fc⑀RI stimulates enhanced SHIP tyrosine phosphorylation leading to association with Shc and Dok. An increase in Dok phosphorylation was observed under co-aggregation conditions, with subsequent increased association with RasGAP (33). Similarly, co-aggregating Fc␥RII with Fc⑀RI on B cells leads to increased tyrosine phosphorylation of Dok associated with negative signaling (50,51). Dok also plays a negative role at multiple steps in T cell signaling (37) and through the ITIM-containing mast cell function-associated antigen (MAFA) receptor on RBL-2H3 cells (52). Overall, our data also support a role for Dok in inhibitory signaling through Fc␥RIIB via its increased association with Grb2 and SHIP.
We observed Dok-Grb2 interaction in non-activated cells, interaction that was decreased at 2 and 5 min in IgE-activated cells and completely absent at 15 min. Thus tyrosine-phosphorylated Dok-Grb2 appears to be important for maintaining Fc⑀RI in its unactivated state and suggests a previously unrecognized gatekeeper role for Dok and Grb2 in human mast cells. This contrasts with a rodent study that found that Dok is not needed for Fc␥RIIB-mediated inhibitory signaling in Dok-deficient mast cells (33), suggesting that another molecule can compensate for loss of Dok or that, as in other signaling systems, there are important differences between rodents and humans.
We have confirmed rodent studies in which Dok was found to interact with SHIP and in which co-aggregation of Fc⑀ with Fc␥ led to increased SHIP phosphorylation (33,53). We observed a high basal tyrosine phosphorylation of Dok-associated SHIP in resting cells that was slightly decreased in IgE-activated (Fc⑀RI-cross-linked) cells. RBL-2H3 cells have also been shown to exhibit high basal tyrosine phosphorylation of SHIP (33). Others have reported an increase in tyrosine phosphorylation of SHIP following Fc⑀RI cross-linking (33). It is possible under these conditions that SHIP is recruited away from Dok and to the Fc⑀RI ␤ chain (33,54,55) to regulate Fc⑀RI signaling through an Fc␥RII-independent mechanism. We are currently examining this hypothesis.
Our data using Dok knock-out BMMCs further point to a role for Dok in the Fc␥RIIB-mediated inhibition of IgE-Fc⑀RI de-granulation. We found that monomeric IgE alone induced degranulation in these cells. Previous studies found similar results in SHIP knock-out mast cells, although the maximal degranulation we observed (Ϸ30%) was lower than that in the SHIP knockouts (Ϸ80%) (31). This suggests that Dok contributes to the process of BMMCs IgE receptor quiescence, but other factors are also needed, probably SHIP. To this end, Ott et al. (33) found Dok was sufficient, but not necessary, for Fc␥RIIB inhibitory signaling in BMMCs from Dok knock-out mice. Taken together, these results suggest that Dok, through its interaction with Grb2 and SHIP, keeps Fc⑀RI in a quiescent state and that this inhibitory complex is maintained under co-aggregating conditions.
We confirm previous studies that implicate SHIP in Fc␥RIIinduced inhibition of Fc⑀RI signaling. The absence of SHIP in mice leads to increased releasability of BMMCs, even during IgE sensitization (31). In human Fc⑀RI studies, Vonakis et al. (56) found a strong negative correlation between the amount of SHIP protein and maximal degranulation in human basophils. Although our results showing reduced Syk phosphorylation support a role for inhibitory signaling by SHP-1 (57), SHIP activation also appears to be important in our human mast cell system. Given that Fc␥RIIB can interact with both SHP-1 and SHIP (58,59), it is still unclear which phosphatase plays the more dominant role in down-regulating Fc⑀RI in human mast cells. It is quite possible that multiple phosphorylation sites in Fc␥RIIB, where phosphatase binding occurs, contribute uniquely to transduction of Fc␥RIIB-mediated inhibitory signals.
Tyrosine phosphorylation of Grb2 and the differential binding of Dok and SHIP may regulate inhibitory function through several different mechanisms. Grb2 may help to localize Dok and/or SHIP to the cell membrane, regulating their inhibitory function. It is well established that SHIP can inhibit Fc⑀RI functional and biochemical responses (60), but it is not as clear how SHIP exerts such effects. Dok tyrosine phosphorylation (or enhanced SHIP tyrosine phosphorylation) may facilitate the recruitment of Dok (through Grb2 as seen in Fig. 7) to the Fc⑀RI-Fc␥RII receptor complex. Indeed, Dok recruitment to Fc⑀RI is enough to inhibit Fc⑀RI-induced signal transduction (30). Thus, because SHIP and Dok can associate with RasGAP under co-aggregating conditions (33,50), Fc⑀RI may be "turned off " by up-regulating Ras GTPase activity. Our findings suggest that the mechanisms that keep Fc⑀RI in its non-activated state are similar to the mechanisms of inhibition through Fc⑀RI-Fc␥RIIB co-aggregation.
Many investigators have sought inhibitors of human Fc⑀RI signaling to block IgE-mediated allergic inflammation. One approach is to block Fc⑀RI signaling by altering the propagation of signals generated from the cross-linked receptors. Strategies based on the specific inhibition of the earliest events in the Fc⑀RI signaling cascade, such as we have shown here using GE2 to co-aggregate Fc⑀RI with Fc␥RII, may ultimately generate new therapies for allergic inflammation. In addition to Fc␥RII, we also demonstrate that CBMCs also express the ITIM-containing receptors CD22 and CD81. We did not detect other ITIM-receptor expression on mast cells but were limited to those which had Abs detecting human isoforms. CD22 is widely known for its ability to down-regulate B cell receptormediated B cell activation (62), whereas CD81 negatively regulates Fc⑀RI responses in rodent mast cells (63). These and other ITIM-containing receptors on human mast cells are also potential novel targets for chimeric bifunctional proteins to "down-regulate" Fc⑀RI function as described here.
In summary, we have demonstrated that IgE activation of cultured human mast cells can be inhibited by co-aggregating Fc⑀RI with Fc␥RII through a novel mechanism involving Grb2, Dok, and SHIP. By achieving this with a human chimeric immunoglobulin, we have sought to turn off allergic responses utilizing the powerful negative regulatory ability of ITIM receptor motifs. The highly specific and avid association of Fc and Fc⑀RI provides a highly specific target by which to deliver the off switch to effector cells of allergy and asthma. This approach can be extended in several ways. Firstly, chimeric proteins capable of binding more that one ITIM receptor (e.g. G-G-E2 or CD81L-G-E2) may be more potent inhibitors. Secondly, it is possible to make chimeric allergen-␥ proteins that will bind to Fc␥RII and indirectly bind to Fc⑀RI via their high affinity interaction with the antigen-specific IgE on the mast cells. Not only would this latter approach serve to inhibit antigen-specific immediate reactivity, but also it should be able to provide a platform for safer and more effective allergen immunotherapy. As such, these improved versions of the chimeric protein described may prove to be promising agents for the treatment of human allergic disorders.