PTPα Activates Lyn and Fyn and Suppresses Hck to Negatively Regulate FcεRI-Dependent Mast Cell Activation and Allergic Responses

Mast cell activation via FcεRI involves activation of the Src family kinases (SFKs) Lyn, Fyn, and Hck that positively or, in the case of Lyn, negatively regulate cellular responses. Little is known of upstream activators of these SFKs in FcεRI-dependent signaling. We investigated the role of receptor protein tyrosine phosphatase (PTP)α, a well-known activator of SFKs in diverse signaling systems, FcεRI-mediated mast cell activation, and IgE-dependent allergic responses in mice. PTPα−/− bone marrow-derived mast cells hyperdegranulate and exhibit increased cytokine and cysteinyl leukotriene secretion, and PTPα−/− mice display enhanced IgE-dependent anaphylaxis. At or proximal to FcεRI, PTPα−/− cells have reduced IgE-dependent activation of Lyn and Fyn, as well as reduced FcεRI and SHIP phosphorylation. In contrast, Hck and Syk activation is enhanced. Syk hyperactivation correlated with its increased phosphorylation at positive regulatory sites and defective phosphorylation at a negative regulatory site. Distal to FcεRI, we observed increased activation of PI3K and MAPK pathways. These findings demonstrate that PTPα activates the FcεRI-coupled kinases Lyn and Fyn and suppresses Hck activity. Furthermore, the findings indicate that hyperactivation of PTPα−/− mast cells and enhanced IgE-dependent allergic responses of PTPα−/− mice are due to the ablated function of PTPα as a critical regulator of Lyn negative signaling.

M ast cells are important effectors of IgE-dependent immune responses that mediate acute and chronic inflammation and allergies (1,2). Mast cell activation is initiated by Ag binding and cross-linking of IgE that is complexed with the high-affinity IgE receptor FcεRI at the mast cell surface. FcεRI engagement and cross-linking triggers intracellular signaling events that culminate in the release, both by degranulation and secretion, of a variety of stored and newly synthesized bioactive compounds that have profound effects on multiple other cells and tissues (3)(4)(5).
The Src family kinases (SFKs) Lyn, Fyn, and Hck are critical FcεRI-proximal regulators of signaling (6). Their individual actions and roles, as well as their functional interrelationships, in FcεRI signaling and mast cell activation have been investigated using bone marrow-derived mast cells (BMMCs) from mice deficient in one or more of these kinases. Lyn exerts both positive and negative regulatory effects that appear to be determined by genetic background and the intensity of mast cell stimulation (7,8). Indeed, Lyn-null BMMCs have been found to have reduced (7)(8)(9)(10), enhanced (7,8,(11)(12)(13), or unaltered (14) Ag-induced degranulation. Lyn associates with the band g-chains of FcεRI and is activated upon receptor clustering, phosphorylating ITAMs in the band g-chains of FcεRI that recruit and promote the activation of other signaling molecules (6,15,16). Among these is the tyrosine kinase Syk that recognizes phospho-ITAMs in FcεRIg. Lyn and Syk phosphorylate several other proteins, such as the transmembrane scaffolding-adaptor proteins linker for activation of T cells (LAT) and non-T cell activation linker, to initiate downstream signaling pathways that stimulate Ca 2+ mobilization and degranulation as well as gene transcription. The positive actions of Lyn in mast cell signaling are manifested in BMMCs that receive low-intensity stimulation or that express a mutant form of Lyn with impaired kinase activity (7,10). A key role of Lyn in negatively regulating mast cell activation was revealed by the hyperactive phenotype of Lyn-null BMMCs and the finding that Lyndeficient mice are prone to allergic disease (7,(11)(12)(13). Lyn-mediated signaling events that are implicated in negatively regulating FcεRIdependent mast cell responses include the phosphorylation of FcεRI b subunits, activation of the inhibitory inositol phosphatase SHIP, and phosphorylation of the transmembrane protein Cbp/PAG (Cterminal Src kinase [Csk]-binding protein/phosphoprotein associated with glycosphingolipid-enriched microdomains) to promote recruitment of the SFK inhibitory kinase Csk (7,12,13). Overall, hyperresponsive Lyn-null BMMCs exhibit enhanced or prolonged activation of Fyn, PI3K, phospholipase C (PLC) g1, and MAPK (12,13,17).
The SFKs Fyn and Hck positively regulate mast cell activation, which is defective in Fyn 2/2 or Hck 2/2 BMMCs (9,11,18). Ag stimulation induces the association of Fyn with FcεRIb, but Fyn does not phosphorylate the receptor (9,11). Both Fyn and Hck phosphorylate the adaptor protein Gab2 to activate PI3K signaling and promote microtubule formation that is important for granule translocation to the plasma membrane (9,11). Hck also positively regulates Syk activation and that of other downstream kinases, such as Btk, Akt, and JNK (9). Evidence of inter-SFK regulation is *Department of Pediatrics, ‡ Department of Pathology and Laboratory Medicine, University of British Columbia; and † Child and Family Research Institute, Vancouver, British Columbia, Canada provided by findings of elevated Lyn activity in Hck 2/2 mast cells together with some Hck-mediated, Lyn-dependent signaling events and by the elevated Fyn expression or activity detected in Lyn 2/2 mast cells (9,(11)(12)(13). This finding suggests that Lyn suppresses positive regulatory actions of Fyn, whereas Hck may function to suppress Lyn albeit without concomitant effects on Fyn activity. In both cases, the SFK-mediated suppression of another SFK is thought to be effected via activation of the SFKnegative regulatory complex Cbp/PAG-Csk.
Although Cbp/PAG-Csk promotes the phosphorylation of the C-terminal tyrosine residue in SFKs to downregulate mast cell SFK activity, little is known of protein tyrosine phosphatases (PTPs) that counter this phosphorylation and maintain or promote SFK activity downstream of FcεRI. Several PTPs with reported activity on SFKs in other signaling systems have been investigated for their roles in FcεRI-mediated signaling using PTP-deficient cells. The receptor PTP CD45 regulates SFK activity in immune-receptor signaling (19), and it is required for BMMC degranulation, cytokine release, and IgE-dependent anaphylactic reactions in mice (20,21). In both unstimulated and Ag-stimulated CD45 2/2 BMMCs, Lyn is hyperphosphorylated at its inhibitory tyrosine residue (21). Variants of the rat basophilic leukemia cell line RBL-2H3 with greatly reduced CD45 expression exhibit defective recruitment of Lyn to FcεRI, reduced FcεRI-associated tyrosine kinase activity and FcεRI tyrosine phosphorylation, and impaired degranulation (22). Thus, defective CD45-mediated regulation of Lyn is a primary event underlying defective FcεRI-mediated responses. BMMCs lacking the nonreceptor PTP SHP-2 exhibit enhanced Ag-induced Lyn activity accompanied by decreased Fyn activation and Fyn-dependent signaling (23). Overall, the lack of SHP-2 did not affect degranulation, but it enhances TNF-a release. The receptor PTPε appears to act via Syk to negatively regulate mast cell activation (24). Ag-stimulated PTPε 2/2 BMMCs showed enhanced degranulation and cytokine release, but unaffected leukotriene secretion. The most FcεRI-proximal alteration detected in these cells was hyperphosphorylation of Syk, whereas Lyn phosphorylation and Fynmediated signaling were normal.
PTPa is a receptor PTP that, like CD45 and the closely related PTPε, has a glycosylated extracellular domain and two tandem catalytic domains, with the majority of activity residing in the membrane proximal domain. In contrast to the mainly hematopoietic expression of CD45 and PTPε, PTPa is widely expressed. PTPa activates receptor-responsive SFKs in diverse signaling pathways (25); it catalyzes the dephosphorylation of the C-terminal inhibitory tyrosine residue of SFKs (26)(27)(28)(29), acting in a manner often coupled to engagement or cross-linking of receptors such as integrins, ligandgated ion channels, and neural cell adhesion molecules (28,(30)(31)(32)(33)(34). In mast cells, PTPa is required for Fyn activation in stem cell factor-dependent c-Kit signaling and mast cell migration (35). Overexpression of Lyn in epithelial Chinese hamster ovary cells stably expressing FcεRI resulted in constitutive phosphorylation of FcεRI because of the dysregulated high basal activity of Lyn, and this was suppressed by co-overexpression of PTPa (36,37). In the absence of Ag stimulation in this system, PTPa acted on the active site phosphotyrosine of Lyn to downregulate its basal kinase activity (37). However, the expressed PTPa reportedly exerted these effects even when present at a level approaching that of endogenous Chinese hamster ovary cell PTPa, whereas the endogenous phosphatase did not, raising the possibility of undefined but critical differences between the heterologous and endogenous PTPa. We investigated the role of PTPa in Ag-stimulated FcεRI signaling and mast cell activation using BMMCs derived from PTPa 2/2 mice. We also determined whether the absence of PTPa affected IgEdependent anaphylactic reactions in mice. We report that PTPa is a negative regulator of these FcεRI-dependent responses. The ablation of PTPa inhibited Ag-stimulated Lyn and Fyn activation while enhancing Hck activation. Despite reduced tyrosine phosphorylation of FcεRIg, Syk activity was elevated in PTPa 2/2 BMMCs, as was the activity of Akt and MAPKs. In conjunction with reduced FcεRIb tyrosine phosphorylation, SHIP phosphorylation was defective, although that of SHP-1 was increased. Our findings indicate that PTPa plays a major role as an upstream regulator of Lyn-mediated negative signaling, revealing a novel regulatory relationship between Lyn and Hck. Overall, PTPa functions as a key FcεRI-proximal phosphatase that initiates and coordinates SFK activities to promote yet limit mast cell activation.

Animals
The 129SvEv PTPa 2/2 mice (29) were backcrossed with C57BL/6 mice for 10 generations. PTPa 2/2 and wild type (WT; PTPa +/+ ) C57BL/6 mice were housed under specific pathogen-free conditions. Animal care and use followed the guidelines of the University of British Columbia and the Canadian Council on Animal Care, and were reviewed and approved by the University of British Columbia.

IgE-mediated passive cutaneous anaphylaxis
To induce passive cutaneous anaphylaxis, mice (five animals per genotype) were injected intradermally with 10 mg IgE (anti-DNP Ab, clone SPE-7) in saline into the left ear and with saline alone into the right ear. After 24 h, mice were challenged by i.v. injection of 1 mg DNP-human serum albumin (HSA) in saline containing 1% Evans Blue (Sigma-Aldrich). Thirty minutes after Ag challenge, mice were sacrificed and the ears were removed. Extravasated Evans blue dye in the ear tissues was extracted with 1 ml of 0.5% Na 2 SO 4 and acetone (3:7) for 24 h, and OD 620 nm was measured and normalized to unit weight. The OD from the saline treated right ear was subtracted from the value for the left ear. A second group of mice (three per genotype) was treated essentially as above. Ear tissues were collected 6 h after DNP-HSA challenge, and histologic analysis was performed as described previously (35). Inflamed ear tissue was visualized and ear thickness was measured using a Leica DM4000 B microscope (Leica Microsystems, Deerfield, IL) and Open Laboratory 4.0.2 software (Improvision, Coventry, England).

IgE-mediated passive systemic anaphylaxis
Mice were sensitized by i.v. injection of 2 mg IgE (anti-DNP Ab, clone SPE-7) in 100 ml saline or treated with saline alone. After 24 h, the mice were challenged i.v. with 500 mg DNP-HSA in 100 ml saline for 1.5 min, and blood was immediately collected by cardiac puncture. Serum histamine concentrations were determined by ELISA (Cayman Chemical, Ann Arbor, MI) according to the manufacturer's specifications.

Kinase assays
Lyn, Fyn, and Hck immunoprecipitates were assayed for kinase activity using the ELISA-based Universal Tyrosine Kinase Assay Kit (GenWay, San Diego, CA) according to the manufacturer's instructions, and activities were defined as a measure of the units of cellular Src (c-Src) kinase activity. To measure Syk kinase activity, immunoprecipitated Syk was assayed using an ELISA-based HTScan Syk Kinase Assay Kit (Cell Signaling Technology) according to the manufacturer's instructions.

Degranulation
Degranulation of BMMCs was measured by determining released bhexosaminidase (38). BMMCs were sensitized with 100 ng/ml anti-DNP IgE and stimulated with varying concentrations of DNP-HSA for 15 min at 37˚C. The cells and supernatant of the degranulated mixture were separated by centrifugation. Cell pellets were solubilized in Tyrodes buffer containing 0.5% Triton X-100 on ice for 15 min followed by centrifugation to remove debris. Supernatants or cell pellet lysates were mixed with 4nitrophenyl N-acetyl-b-D-glucosaminide (Sigma-Aldrich), a substrate for b-hexosaminidase, in sodium citrate (pH 4.5) for 15 min at 37˚C. The enzymatic reaction was stopped by the addition of 0.1 M sodium carbonate (pH 10), and the absorbance at 405 nm was measured. Degranulation was expressed as a percentage of b-hexosaminidase released (absorbance of supernatant divided by absorbance of supernatant plus absorbance of pelleted cell lysate). To determine the amount of phosphatidylserine externalization as a measure of degranulation, BMMCs were degranulated essentially as above, followed by labeling with PE annexin V and 7AAD according to the manufacturer's specifications, and assessed using FACS.

Mast cell mediator release
BMMCs were incubated for 5 h with anti-DNP IgE and stimulated with DNP-HSA for 15 h. Culture supernatants were analyzed for IL-2, IL-6, and TNF-a using ELISA kits (R&D Systems, Minneapolis, MN). Alternatively, supernatants were collected after 1 h of stimulation and assessed for cysteinyl leukotrienes using an EIA kit (Cayman Chemical).

Data analysis
Data were statistically analyzed using the Student t test.

PTPa deficiency leads to Fc«RI-dependent hyperdegranulation and augmented cytokine production in BMMCs
To determine whether PTPa regulates FcεRI-dependent release of bioactive mediators from mast cells, we measured mast cell de-FIGURE 1. PTPa deficiency leads to enhanced degranulation and cytokine production. A, WT (PTPa +/+ ) and PTPa 2/2 BMMCs (4-6 wk culture) were analyzed for c-Kit and FcεRI expression by FACS to confirm purity prior to use. Representative FACS graphs are shown. For activation, the cells were sensitized overnight with anti DNP-IgE mAb and challenged with DNP-HSA for 15 min. B, Released b-hexosaminidase was measured. The inset shows total (released plus nonreleased) b-hexosaminidase of the cells (WT, +/+ ; PTPa 2/2 , 2/2 ) in arbitrary absorbance units. C, Cells were stained with annexin V and analyzed by FACS, and the percentage of phosphatidylserine externalization was determined. A representative FACS graph showing annexin V stained scatter for 10 ng/ml DNP-HSA stimulation is shown. Numbers in the bottom right quadrant show the percentage of annexin V-positive cells. D-G, Cytokine release was measured from cells after stimulation with DNP-HSA for 15 h; IL-2 (D), IL-6 (E), and TNF-a (F). G, Cysteinyl leukotriene release was measured after 1 h of stimulation. Data represent the mean 6 SD from three separate experiments. Asterisks indicate a significant difference (p , 0.05) between comparative data from both genotypes. granulation and cytokine production from BMMCs established from WT and PTPa 2/2 mice. Both types of BMMCs exhibited equivalent expression of the mature mast cell markers c-Kit and FcεRI (Fig. 1A). FcεRI-mediated degranulation of PTPa 2/2 mast cells was markedly enhanced (up to 2-fold) compared with WT mast cells (Fig. 1B) despite the equivalent b-hexosaminidase content of both types of BMMCs (Fig. 1B, inset). To confirm the observed hyperdegranulation phenotype of PTPa 2/2 mast cells, annexin V-positive phosphatidylserine on the externalized membrane regions of degranulated mast cells was assessed. Degranulated PTPa 2/2 mast cells showed significantly higher phosphatidylserine externalization compared with WT cells (Fig. 1C). Activated mast cells not only release cytokines by degranulation, but also synthesize and secrete a variety of cytokines in late phase responses. We therefore determined FcεRI-mediated cytokine release of WT and PTPa 2/2 mast cells. Significantly elevated levels of IL-2, IL-6, TNF-a, and cysteinyl leukotrienes were released from PTPa-deficient BMMCs (Fig. 1D-G). Collectively, these results demonstrate that PTPa is a negative regulator of FcεRI-dependent BMMC degranulation and cytokine release.
To confirm that the observed hyperactivation phenotype of PTPa 2/2 BMMCs was indeed due to the absence of PTPa, PTPa was re-expressed in the PTPa-null BMMCs by retroviral (i.e., MSCV) infection, to a level that was 86 6 11% (n = 3) of that in WT BMMCs ( Fig. 2A). PTPa re-expression effectively abolished the ability of the BMMCs to hyperdegranulate, because b-hexosaminidase release induced by all concentrations of Ag tested was reduced to levels similar to those of WT BMMCs (Fig. 2B). Likewise, the re-expression of PTPa resulted in the loss of the abilities of these mutant cells to release elevated levels of IL-2 and TNF-a (Fig. 2C). PTPa 2/2 BMMCs infected with MSCV carrying a control (empty) plasmid exhibited degranulation and cytokine release characteristic of the parental PTPa 2/2 BMMCs (Fig. 2B, 2C).

PTPa-deficient mice show significantly enhanced anaphylactic responses
Anaphylaxis is a profound allergic reaction initiated by allergeninduced cross-linking of specific IgE bound to FcεRI. IgE/FcεRI-stimulated mediator release from mast cells is a major determinant of anaphylaxis (2,39). Because PTPa-deficient mast cells exhibit FcεRI-dependent hyperdegranulation ex vivo, we questioned whether PTPa plays a role in FcεRI-mediated anaphylaxis in mice. Passive cutaneous anaphylaxis (PCA) was examined upon i.v. challenge with Ag (DNP-HSA in 0.5% Evans blue dye) of sensitized WT and PTPa-deficient mice. As shown in Fig. 3A, mice lacking PTPa demonstrated enhanced PCA as reflected by the significantly elevated Evans blue dye extravasation that specifically occurred in IgE-sensitized ear tissue. We qualitatively assessed IgE-mediated local inflammation responses by measuring inflamed ear tissue thickness and mast cell numbers in inflamed tissue 6 h after Ag challenge. PTPa 2/2 mice generally had thicker inflamed ears than WT mice (knockout, 839 6 49 mm; WT, 795 6 73 mm; n = 15 microscopic fields). Strikingly, we also observed an unusual skin inflammatory response of highly thickened regions of inflamed epidermis that was present only in IgE-sensitized PTPa 2/2 ears (Fig. 3B, bottom right panel). Passive systemic anaphylaxis (PSA) was assessed in mice sensitized by i.v. injection of DNP-IgE or control saline and challenged 24 h later with an i.v. injection of DNP-HSA. Serum histamine levels, a measure of PSA, were significantly elevated in IgE-sensitized PTPa 2/2 mice (Fig. 3C). Our data demonstrate that PTPa negatively regulates IgE-and FcεRIdependent PCA and PSA, consistent with the ex vivo hyperdegranulation of cultured BMMCs. PTPa 2/2 mice have reduced mast cell numbers in many tissues, including the ear (35); this suggests that the enhanced PCA response, and perhaps other anaphylactic responses in these animals, results from an even higher extent of activation on a per cell basis.

PTPa differentially regulates Fc«RI-mediated activation of SFKs
PTPa is well established as an important regulator of SFKs, acting to dephosphorylate their inhibitory C-terminal tyrosine residue and activate them (25). Cross-linking of FcεRI-bound IgE leads to the activation of the SFKs Fyn, Lyn, and Hck-all critical receptor-immediate effectors of mast cell activation (6). We there- fore investigated the effects of PTPa ablation on FcεRI-dependent activation of these SFKs. We used phosphospecific Abs to Src phosphoTyr 416 to analyze active SFKs, Abs to Src phosphoTyr 527 to analyze inactive SFKs, and enzymatic assays to quantify kinase activity. Immunoblotting and kinase assay analysis revealed little to no Lyn activation in stimulated PTPa 2/2 BMMCs, resulting in significantly reduced Lyn activity for up to 15 min poststimulation in PTPa 2/2 BMMCs compared with WT BMMCs (Fig.  4A). We also found a lack of Fyn activation in stimulated PTPa 2/2 BMMCs, as indicated by the absence of Fyn Tyr528 dephosphorylation and reduced kinase activity of Fyn in stimulated PTPa 2/2 mast cells compared with WT cells (Fig. 4B). In contrast, Hck activation was evident in IgE cross-linked PTPa 2/2 BMMCs and was higher than that detected in WT cells, as shown by enhanced phosphorylation of Hck at its Tyr 416 -equivalent residue and by its significantly increased (1.5 and 5 min) and prolonged (15 min) activation in PTPa 2/2 cells (Fig. 4C). These analyses reveal unexpected differential abilities of PTPa to regulate FcεRI-dependent SFK activities, with PTPa positively regulating Fyn and Lyn activation but negatively regulating Hck activation. PTPa functions as an activator of SFKs in multiple signaling systems, indicating that reduced Lyn and Fyn activities in PTPa 2/2 BMMCs reflect the loss of PTPa-catalyzed dephosphorylation and activation of these kinases. In contrast, it is FIGURE 3. Lack of PTPa significantly elevates anaphylaxis in mice. Passive cutaneous anaphylaxis was induced in WT and PTPa 2/2 mice. A, Thirty minutes after Ag challenge, ears were removed and extravasated Evans blue dye was quantified and normalized for the unit weight (n = 5 animals per genotype). B, Six hours after Ag challenge, ear tissues were assessed for tissue inflammation. One of three representative micrographs per condition is shown (n = 3 animals per genotype). Characteristic thickening of the epidermis of Ag-IgE-treated PTPa 2/2 mouse ears is marked by the arrow. Scale bar, 100 mm. C, Passive systemic anaphylaxis was induced in mice (n = 5 animals per genotype). Blood was collected 1.5 min after Ag challenge, and serum histamine concentrations were determined by ELISA. The graph represents mean 6 SD. Asterisks indicate a significant difference (p , 0.05) between comparative data from animals of each genotype. , and Hck (C) were prepared from lysates of WT and PTPa 2/2 BMMCs that were not stimulated (0) or stimulated with 100 ng/ml DNP-HSA for the indicated times. Immunoprecipitates were immunoblotted with anti-phospho Tyr 527 -Src Ab or anti-phospho Tyr 416 -Src Ab (top panels) and with anti-Fyn, anti-Lyn and anti-Hck Ab as appropriate (bottom panels). A representative set from three similar sets of immunoblots is shown. The corresponding graphs below show the densitometric quantification (mean 6 SD) of immunoblots from three independent experiments. The graphs to the right show the activity of each immunoprecipitated SFK, as determined by in vitro kinase assays. Kinase activity was calculated relative to the standard kinase activity of c-Src. Data represent the mean 6 SD calculated from duplicate assays of each of three independent immunopreciptates. The asterisks indicate a significant difference (p , 0.05) between comparative data from each genotype. D, Cbp/PAG was immunoprecipitated and probed with phosphotyrosine (4G10, top panel), and stripped and reprobed with Cbp/PAG and Csk Abs (bottom panels). The graph shows the densitometric quantification of the amount of Csk that coimmunoprecipitated per unit Cbp/PAG (mean 6 SD) from three independent experiments. n.s., no significant difference (p . 0.05).
likely that the enhanced Hck activity in these cells results from the loss of indirect PTPa-mediated signaling events that normally suppress Hck. One possibility is that reduced Lyn activation results in reduced Cbp/PAG-mediated membrane recruitment of the SFK inhibitor Csk, although we did not detect significantly reduced Cbp/PAG-associated Csk in PTPa 2/2 BMMCs (Fig. 4D).

PTPa positively regulates Fc«RI receptor phosphorylation and negatively regulates Fc«RI-dependent MAPK and PI3K pathways
Overall protein tyrosine phosphorylation was enhanced in Ag-stimulated PTPa 2/2 mast cells (Fig. 5A). To further determine how the absence of PTPa affected specific steps in FcεRImediated signaling to result in mast cell hyperactivation, we examined the early event of FcεRI phosphorylation and the later activation of downstream molecules in the PI3K and MAPK signaling pathways. Lyn plays a critical role in the phosphorylation of the FcεRI-b and FcεRI-g subunits to initiate subsequent signaling (3,7), and Hck can also regulate FcεRI-b and FcεRI-g phosphorylation (9). Fyn does not phosphorylate FcεRI (11,18). The FcεRI-g subunit was immunoprecipitated from WT and PTPa 2/2 BMMCs and probed for tyrosine phosphorylation (Fig.  5A) and associated phosphotyrosyl b subunit (Fig. 4B, 4C). We observed a large decrease (∼70%) in Ag-stimulated phosphorylation of the FcεRI-g subunit (Fig. 5B) and reduced phosphorylation of the associated b subunit (80% of WT; n = 2; Fig 5C), in PTPa-deficient mast cells compared with WT cells. These results are in accord with the decreased Lyn activity in PTPa-null BMMCs; they also are consistent with the increased Hck activity in PTPa 2/2 mast cells, because Hck 2/2 BMMCs exhibit enhanced FcεRI g and b phosphorylation that may be due to the proposed role of Hck as a negative regulator of Lyn and Lynmediated phosphorylation events (9).
Despite reduced FcεRI tyrosine phosphorylation, we detected hyperphosphorylation of Akt (Fig. 5D), ERK1/2 (Fig. 5E), p38 (Fig.  5F), and JNK (Fig. 5G) in PTPa 2/2 cells, which is indicative of the increased activation of PI3K and MAPK signaling. These findings demonstrate that, despite the action of PTPa to positively regulate FcεRI receptor phosphorylation, it nevertheless acts as a negative regulator of several key downstream signaling molecules.
Increased activity of the Fc«RI-interacting kinase Syk in PTPa 2/2 mast cells Lyn-mediated phosphorylation of the g and b subunits of FcεRI results in the recruitment of positive and negative regulators of signaling. The tyrosine kinase Syk becomes activated upon interaction with phosphorylated ITAMs of FcεRIg, subsequently phosphorylating diverse target molecules to elicit responses that include mast cell degranulation and cytokine release (40). Interestingly, we found that Syk phosphorylation at tyrosine 525/ 526, reflecting Syk activation, was significantly enhanced in PTPa 2/2 BMMCs at 1.5 min after stimulation, and it was equivalent to that in WT BMMCs by 5 and 15 min after stimulation (Fig. 6A). In vitro assays of immunoprecipitated Syk Densitometric quantification (mean 6 SD) of immunoblots from three experiments is shown in the graph. Asterisks indicate a significant difference (p , 0.05) between comparative data from the two genotypes. C, In other experiments, FcεRIg immunoprecipitates were probed as in B and reprobed for FcεRIg (top panels) and FcεRIb (bottom panels; n = 2). D, Cell lysates were probed for phospho-Ser473 Akt and anti-Akt, (E) phospho-ERK1/2 and anti-ERK, (F) phospho-p38 MAPK and anti-p38 MAPK, and (G) phospho-JNK1/2 and anti-JNK1/2, as indicated (n = 3). Representative immunoblots of three similar independent experiments are shown, and the graphs depict the mean 6 SD of quantified data from three experiments. Asterisks indicate a significant difference (p , 0.05) between comparative data from the two genotypes. demonstrated significantly elevated Syk activity in PTPa 2/2 BMMCs, compared with WT BMMCs, that was evident for up to 15 min after stimulation (Fig. 6B). These results demonstrate that Syk activation is PTPa dependent, and the enhanced activation of Syk in PTPa 2/2 mast cells is an early postreceptor signaling event that correlates with the enhanced degranulation and cytokine secretion responses that we observed in PTPa 2/2 BMMCs.
Syk binding to phospho-FcεRIg is required for Syk activation. Because stimulated PTPa 2/2 BMMCs have reduced FcεRIg phosphorylation, but have increased Syk activity, we investigated the interaction of Syk with FcεRIg in PTPa 2/2 and WT cells. Immunoprecipitations of FcεRIg performed using Abs to the g subunit did not detect any bound Syk in WT mast cells (data not shown). Therefore, we used anti-IgE Abs to immunoprecipitate activated FcεRI and then probed for FcεRIg and associated Syk. Although Syk association with FcεRI increased upon Ag stimulation of WT BMMCs, no Ag-dependent increase was observed in PTPa 2/2 BMMCs (Fig. 6C). This finding suggests that Syk is hyperactivated in PTPa-null mast cells via a mechanism other than that involving enhanced Syk binding to FcεRI.
We determined Syk phosphorylation at Tyr317, a site in the linker region that inhibits Syk function (41). Ag-stimulated phosphorylation of Tyr317 was not detectable in PTPa 2/2 mast cells, whereas it increased by ∼30% in WT cells (Fig. 6D). Therefore, the enhanced Syk activity in PTPa-null BMMCs could result from the combined effects of the absence of inhibitory phosphorylation and gain in positive regulatory phosphorylation.
The scaffolding protein Gab2 is a positive regulator of FcεRImediated mast cell degranulation and anaphylaxis (42). Previous work has shown that Syk can phosphorylate Gab2 (43). Importantly, Fyn and Hck also mediate Gab2 phosphorylation to regulate PI3K activation, microtubule formation, and degranulation (9,11). In parallel with the enhanced activation of Syk and Hck, we observed hyperphosphorylation of Gab2 in PTPa 2/2 mast cells at 1.5 min, but not at times of 5 and 15 min after stimulation (Fig.  6E). Therefore, PTPa functions as a negative regulator of FcεRIdependent activation of Syk and Gab2.

Regulation of Fc«RI-associated phosphatases by PTPa
Lyn is associated with and phosphorylates the FcεRI-b subunit to mediate negative signaling that includes recruitment and phosphorylation of the inositol phosphatase SHIP and the protein tyrosine phosphatase SHP-1 (7,13,(44)(45)(46). In PTPa 2/2 BMMCs, Ag-stimulated phosphorylation of SHIP was inhibited while that of SHP-1 was increased (Fig. 7A, 7B).

Discussion
We have used PTPa-deficient BMMCs and mice to define roles of PTPa as a negative regulator of FcεRI-mediated mast cell activation and IgE-dependent anaphylaxis. Ablation of PTPa has profound effects on the Ag-stimulated activity of three SFKs that critically mediate early events in FcεRI signaling. In PTPa-null BMMCs, Ag-stimulated activation of the SFKs Lyn and Fyn is abrogated, whereas Hck activation is enhanced. In conjunction with the well-established actions of PTPa to activate SFKs in multiple signaling systems and its demonstrated phosphatase activity toward Fyn (25,47), these findings indicate that PTPa acts as an Ag-stimulated Lyn-Fyn phosphatase and a critical positive regulator of these SFKs in FcεRI signaling, in addition to acting as a negative regulator of Hck (Fig. 8).
The resemblance of the situation of enhanced Hck activity and reduced Lyn activity in PTPa 2/2 BMMCs to the enhanced Fyn activity in Lyn 2/2 BMMCs (12, 13) supports the likelihood that enhanced activation of Hck is due to a similar mechanism involving its reduced inhibition, in this case possibly mediated through the loss of PTPa-induced Lyn activation. The elevated Hck activation did not appear to be mediated through a reduction in Lyn-promoted Cbp/PAG-Csk signaling to suppress Hck, and it requires further investigation. It will be of interest to determine whether Hck activity is elevated in Lyn 2/2 and Lyn WeeB mast cells that respectively have no Lyn protein or express mutant Lyn with severely reduced activity. A potential for bidirectional signaling between Lyn and Hck is suggested by the observation that Hck 2/2 mast cells possess elevated Lyn activity, indicating that Hck may also inhibit Lyn (9). Because Lyn can inhibit Fyn (12,13), PTPa may additionally promote Lyn-mediated inhibition of Fyn, although apparently the direct activation of Fyn by PTPa can effectively overcome this indirect mechanism of inhibition. Obviously, inter-SFK regulatory actions are complex and are determined by the overall SFK expression levels and activities, with the latter controlled both directly and indirectly by PTPa (Fig. 8).
Among Lyn-, Fyn-, or Hck-deficient BMMCs, only Lyn 2/2 cells exhibit a hyperactivation phenotype of enhanced degranulation, cytokine release, or both (7,(11)(12)(13)17). The SFK activity profile in Ag-stimulated PTPa 2/2 BMMCs suggests that the loss of Lynmediated negative regulatory signals, which in part is manifested by a consequent gain of Hck-mediated positive regulatory signals, underlies the hyperactivated phenotype of these mast cells (Fig.  8). A role for elevated Hck activity in PTPa 2/2 mast cell hyperactivation is indicated by findings that the Lyn 2/2 hyperactivation phenotype is supported by catalytically active Hck, as double-deficient Lyn 2/2 Hck 2/2 BMMCs show reduced hyperactivation relative to Lyn 2/2 cells that is restored by the introduction of active Hck into the double-deficient cells (9). Similarly, the amount of Fyn expressed in Lyn 2/2 BMMCs influences degranulation, because the Lyn 2/2 hyperdegranulation phenotype is lost in double-deficient Lyn 2/2 Fyn 2/2 BMMCs or in Lyn 2/2 BMMCs with reduced Fyn expression, and the degranulation of Lyn 2/2 cells is increased by expression of Fyn (8,12). The abilities of catalytically active Hck and Fyn to promote and enhance the manifestation of the Lyn 2/2 hyperactivation  phenotype, together with the overlap of some Hck and Fyn signaling activities (9,11), suggest that the increased Hck activation in PTPa 2/2 BMMCs at least partially contributes to PTPa 2/2 BMMC hyperactivation through positive regulatory actions that are Hck-specific and by compensating for some of the abrogated Fyn-mediated positive signaling effects.
Another aspect of Lyn-negative regulatory signaling involves the activation of molecules such as the inositol phosphatase SHIP (7,13), which acts to dephosphorylate PIP3 and limit PI3K signaling (44). SHIP activation appears to be defective in the absence of PTPa, because reduced phosphorylation of SHIP and enhanced phosphorylation of the PI3K target Akt occur in PTPa 2/2 BMMCs. Like SHIP, the PTP SHP-1 undergoes FcεRI-mediated and Lyn-dependent phosphorylation and activation (7,46,48). SHP-1 is reported to dephosphorylate various signaling molecules, including FcεRIb/g, Syk, LAT, and SLP-76, and to have positive and negative effects on mast cell signaling and activation (49,50). We detected only a small increase in SHP-1 tyrosine phosphorylation in WT BMMCs, but a greatly enhanced phosphorylation of SHP-1 was apparent in Ag-stimulated PTPa 2/2 BMMCs. This finding suggests that SHP-1 catalytic activity and its regulatory actions are unexpectedly enhanced in the absence of PTPa. Nevertheless, this enhancement is insufficient to exert a net inhibition of PTPa 2/2 BMMC activation. However, SHP-1 targets remain to be clarified, and because a distinct positive role of SHP-1 in FcεRI signaling may involve its function as an adaptor that facilitates SLP-76 and PLCg association (50), the significance of enhanced SHP-1 phosphorylation in PTPa-null cells is difficult to interpret. SHP-1 autodephosphorylation has been reported in RBL-2H3 basophilic leukemia cells (48). An additional mode of regulation is indicated by our finding that PTPa directly or indirectly limits Ag-stimulated SHP-1 phosphorylation.
A key positive action of Lyn in FcεRI-mediated signaling is the phosphorylation of ITAMs in FcεRI-g that recruit and activate Syk. In keeping with this role, Lyn 2/2 mast cells typically exhibit reduced Syk phosphorylation and activation (7)(8)(9)11). In Lyn WeeB mast cells, very low Lyn activity is sufficient for positive signaling events that are indicative of intact Syk activation, such as the phosphorylation of the Syk substrate LAT (10). In PTPa 2/2 BMMCs, Ag-stimulated Lyn activity and phosphorylation of FcεRI-g are reduced, but still detectable, and are sufficient to support Syk activation. In fact, Syk is hyperactivated in these mast cells, despite its reduced association with FcεRI-g after Ag stimulation. This finding appears in part to be a result of early enhanced phosphorylation of Syk at the Tyr525/6 sites associated with kinase activation and signaling function in addition to the loss of Ag-dependent inhibitory phosphorylation at Tyr317. These findings suggest that the elimination of PTPa-dependent signaling, certainly that involving Lyn activation, is manifested through the loss of negative but not positive regulation of Syk.
A recent study of Lyn WeeB BMMCs found that this mutant form of Lyn with greatly reduced basal and Ag-stimulated catalytic activity was sufficient for the positive regulatory actions of this SFK in mast cell activation (10). Our study is in agreement with and extends this finding, because it demonstrates that Lyn-positive signaling is functional in PTPa 2/2 mast cells with WT Lyn expression and basal activity, but with minimal Lyn activation. We also identify PTPa as a critical upstream initiator of the majority of Ag-and FcεRI-induced Lyn activation that drives negative regulatory signaling. Furthermore, the ex vivo hyperactive phenotype of PTPa 2/2 mast cells is reflected by enhanced PSA and PCA in mice, indicating that PTPa limits FcεRI-initiated anaphylactic allergic reactions. Further investigation of the mechanisms by which Ag engagement of FcεRI promotes PTPa to function in what appears to be one of the earliest enzymatic reactions that triggers subsequent signaling, and of whether PTPa targets a discrete population of Lyn while another is inaccessible to PTPa, will be required to explore the utility of therapeutically manipulating PTPa to control allergic disease.