SH2 Domain-Containing Phosphatase-2 Protein-Tyrosine Phosphatase Promotes FcεRI-Induced Activation of Fyn and Erk Pathways Leading to TNFα Release from Bone Marrow-Derived Mast Cells¹

Mast cells are key effectors of IgE-hypersensitivity disorders and express high levels of FcεRI that bind IgE, rendering the cells sensitive to particular Ags. Although mast cells play a protective role in bacterial and parasitic infections, dysregulation of mast cells has been implicated in the development of many immune disorders, including type I hypersensitivity reactions, asthma, and multiple sclerosis (1). Binding of multivalent Ags leads to FcεRI clustering, and signaling events promoting degranulation and the production of numerous cytokines, chemokines, and eicosanoids (2). FcεRI signaling is initiated by the Src family kinase (SFK)⁴ Lyn that phosphorylates ITAMs in β- and γ-chains that serve as recruitment sites for SH2 domain-containing adaptor proteins, including protein-tyrosine kinases (PTKs), PI3K, and phosphatases (3). In addition, downstream signaling leads to tyrosine phosphorylation- and phosphoinositide-based recruitment of many signaling proteins to transmembrane adaptors (e.g., linker for activation of T cells (LAT), non-T cell activation linker (NTAL), or LAT2) and pleckstrin homology domain-containing adaptors (e.g., SLP-76, Grb2-associated binder 2 (Gab2)) (4). Mast cell degranulation requires calcium mobilization through Syk-mediated phosphorylation of phospholipase C-γ (PLCγ) following recruitment to LAT, along with a calcium-independent pathway driving microtubule-dependent granule translocation to the plasma membrane through Gab2 adaptor and activation of Fyn kinase and RhoA (5). FcεRI signaling also leads to activation of Erk, Jnk, and p38 MAPKs; Akt/PKB; and IκB kinases (IKK) that promote activation of transcription factors and up-regulation of numerous cytokine and chemokine genes (3).

Tight control of the extent and duration of FcεRI signaling and mast cell activation involves the recruitment and activation of inhibitors of signaling. These include Lyn PTK, which phosphorylates and activates SHIP (a 5′ inositol phosphatase) as well as Src homology region 2 domain-containing phosphatase 1 (SHP1) protein-tyrosine phosphatase (PTP) (6, 7), which both function as inhibitors of mast cell activation (8, 9). FcεRI-induced phosphorylation of ITIM-bearing receptors (e.g., FcγRIIb, PECAM-1, gp49B1, mast cell function-associated Ag (MAFA)) represents one recruitment mechanism for these phosphatases, as well as possible direct binding to phosphorylated FcεRI subunits (reviewed in Refs. 3, 10). FcεRI signaling also leads to recruitment of SH2 domain-containing phosphatase-2 (SHP2) (PTPN11) phosphatase to FcεRI β-subunit (11), Gab2 adaptor protein (12), PECAM-1 (13, 14), and MAFA (15). SHP2 interaction with these ligands likely contributes to activation of SHP2 phosphatase activity, which was previously reported following FcεRI aggregation in mast cells (16, 17). In RBL-2H3 rat mucosal mast cells, clustering of MAFA led to SHP2 recruitment and reduced Syk kinase activation (18), suggesting that Syk is a potential substrate of SHP2 following membrane recruitment and activation (3). In addition, clustering of Thy-1 in RBL-2H3 cells has been shown to trigger SHP2 recruitment and activation independent of FcεRI (16).

Activation of SHP2 normally requires interactions of its SH2 domains with phosphorylated ligands to prevent autoinhibition, and recently activating mutations in the shp2 (ptpn11) gene have been identified in Noonan syndrome and several cancers (reviewed in Ref. 19). These mutations disrupt the autoinhibitory conformation of SHP2 that involves interactions between the N-terminal SH2 and the PTP domain, resulting in constitutive activation of SHP2. Expression of SHP2 mutants found in juvenile myelomonocytic leukemia in mouse bone marrow was sufficient to induce juvenile myelomonocytic leukemia-like disease in mice, and correlated with elevated Erk, Akt, and Stat3 activation (20). More recently, knock-in mice expressing the SHP2^D61Y leukemia-associated mutant in all hematopoietic cells were shown to develop a fatal myeloproliferative disorder, with elevated cytokine-induced activation of Erk and Akt kinases and colony formation (21). Expression of SHP2^D61Y in cardiac endocardium was recently shown to enhance endocardial-mesenchymal transformation and was associated with increased Erk activation (22). These results are consistent with the role of SHP2 as a positive regulator of the Ras/Erk pathway downstream of the epidermal growth factor receptor (EGFR). In this pathway, SHP2 may promote Ras signaling by antagonizing the recruitment of Ras GTPase-activating protein (RasGAP) via dephosphorylation of RasGAP-binding sites within EGFR and the Gab1 adaptor protein (reviewed in Ref. 23). Another proposed mechanism involves SHP2 dephosphorylating the transmembrane adaptor Cbp/PAG to limit recruitment of the C-terminal Src kinase (Csk), which negatively regulates activities of SFKs (24).

Deciphering the role of SHP2 in a specific cell type has been hampered by embryonic lethality in shp2-null mice (25, 26) and the key role of SHP2 in myeloid differentiation (27). However, development of conditional (LoxP-targeted) shp2^flox alleles for exon 4 (28) or exon 11 (26) has now made it possible to uncover the functions of SHP2 in adult mice using tissue-specific or regulated Cre recombinase-expressing transgenic mice. These studies have implicated SHP2 in regulating metabolism (28), obesity, and diabetes (29); liver regeneration (30); and insulin sensitivity and cardiovascular protection (31). Recently, Gu and coworkers (33) used 4-hydroxytamoxifen (4OH-TM)-regulated Cre-modified estrogen receptor fusion protein (CreER*) to inactivate shp2 in mature bone marrow-derived mast cells (BMMCs) and showed that SHP2 is a positive regulator of stem cell factor (SCF)-induced proliferation of BMMCs (33). Likewise, hematopoietic stem cells/progenitor cells expressing SHP2^D61Y were recently shown to display elevated SCF-induced activation of Erk and Akt kinases and elevated colony formation (21).

In this study, we investigated the effect of SHP2 depletion on FcεRI signaling in BMMCs. Using exon 4-targeted shp2^fl/fl mice (28) crossed with transgenic mice expressing CreER* (32), we obtained mature BMMCs before shp2 inactivation following 4OH-TM treatment. Depletion of SHP2 resulted in extended FcεRI-evoked phosphorylation of Lyn and Syk PTKs, compared with control BMMCs. Although SHP2 depletion had no effect on degranulation, it did result in less activation of Erk and Jnk MAPKs. SHP2-deficient BMMCs displayed elevated phosphorylation of Fyn on the inhibitory pY531 residue, and reduced Fyn PTK-directed signaling to Gab2/Akt and IKK/NF-κB pathways and TNF-α release in SHP2 depleted cells compared with control. This study provides new insights into the role of SHP2 as a positive effector of FcεRI signaling and cytokine production.

Transgenic mice harboring LoxP targeting of shp2 exon 4 (shp2^fl/fl) have been described previously (28) and were maintained on a C57BL/6 background. Transgenic mice that ubiquitously express a Cre recombinase-modified estrogen receptor fusion (Tg^CreER*) were described previously (32), and were obtained from The Jackson Laboratory (Strain name: B6.Cg-Tg(cre/Esr1)5Amc/J) on a mixed (C57BL/6 × CBA) background. The shp2^fl/fl and the Tg^CreER* transgenic mouse lines were interbred at Queen’s Animal Care Services to generate congenic littermates that were shp2^fl/fl (control) or shp2^fl/fl:Tg^CreER* genotypes. All mouse studies were approved by the Queen’s University Animal Care Committee.

Noncommercial Abs used in this study included: anti-Syk and anti-Lyn (provided by Joan Brugge; Harvard University, Boston, MA). Commercial Abs used include: rabbit anti-SHP2 (sc-280; SCBT), mouse anti-phospho-tyrosine PY99 (sc-7020; SCBT), rabbit anti-Akt 1/2 (sc-8312; SCBT), rabbit anti-Erk (sc-94; SCBT), mouse anti-phospho-Erk (sc-7383; SCBT), mouse anti-Fyn (sc-434; SCBT), rabbit anti-phospho-Y416 Src (no. 2101, CST), rabbit anti-phospho-Y527 Src (no. 2105, CST), rabbit anti-phospho-Y452 Gab2 (no. 3882, Y441 in mouse, CST), rabbit anti-phospho-S473-Akt (no. 4058, CST), rabbit anti-p38 (no. 9212, CST), rabbit anti-phospho-T180/Y182-p38 (no. 9211, CST), mouse anti-phospho-T183/Y185 Jnk (no. 9255, CST), rabbit anti-phospho-Y783 PLCγ1 (no. 2821; CST), rabbit anti-PLCγ1 (sc-81; SCBT), rabbit anti-phospho-Y1217 PLCγ2 (no. 3871, CST), rabbit anti-PLCγ2 (sc-407; SCBT), rabbit anti-phospho-Y352 Syk (pY346 in mouse; no. 2701, CST), rabbit anti-phospho-S180/S181 IKKα/IKKβ (no. 2681, CST); rabbit anti-IKK (no. 2682, CST), rabbit anti-IκB (no. 9242; CST), rabbit anti-Gab2 (no. 06-967, Upstate Biotechnology), and rabbit anti-phospho-Y507 Lyn (clone EP504Y, Epitomics).

BMMC cultures were generated in a manner similar to what has been previously described (14, 34). Femurs were isolated from shp2^fl/fl and shp2^fl/fl:Tg^CreER* mice, flushed with BMMC growth medium (IMDM, 10% (v/v) FBS, 1% (v/v) antimicrobial-antimycotic solution (Invitrogen), 1 mM sodium pyruvate (Invitrogen), 1% (v/v) nonessential amino acids (Invitrogen), 5% (v/v) conditioned medium from X63-IL-3 cells, 5% (v/v) conditioned medium from HEK293-SCF cells, and 50 μM α-monothioylglycolate (Sigma-Aldrich)) and bone marrow cells were cultured for 4 wk to generate BMMC cultures. Homogeneity of BMMC cultures and cell maturity was confirmed by sensitizing the BMMCs with 20% volume anti-DNP IgE conditioned medium (SPE7 clone) and detected using anti-IgE-FITC (Southern Biotechnology Associates) and anti-Kit-PE (Caltag Laboratories). Gating for nonspecific binding was performed using isotype controls: rat anti-IgG1-FITC (Caltag Laboratories) and rat anti-IgG2b-PE (Caltag Laboratories) to stain the cells. Subsequently, the cells were analyzed by flow cytometry (EPICS Altra HSS; Queen’s University Cancer Research Institute) to ensure comparable maturity between BMMC cultures (>90% FcεRI⁺/Kit⁺ cells).

Mature BMMCs from shp2^fl/fl and shp2^fl/fl:Tg^CreER* genotypes were treated with BMMC growth medium supplemented with 100 nM 4OH-TM for 3 days to generate wild-type (WT) and knockout (KO) cultures, respectively. SHP2 and Erk immunoblots (IB) were performed to confirm a reduction in SHP2 protein levels (>80% in KO compared with WT) was achieved, as measured by densitometry. The deletion of shp2 exon 4 was also confirmed by PCR from genomic DNA from BMMCs (data not shown), as described in Ref. 28 . BMMCs (10⁷ cells/time point) were then starved of IL-3/SCF and sensitized with 20% (v/v) anti-DNP-IgE conditioned medium (SPE7 clone; ≈1 μg/ml) for 6 h, washed in Tyrode’s buffer (10 mM HEPES (pH 7.4), 130 mM NaCl, 5 mM KCl, 1.4 mM CaCl, 1 mM MgCl, 5.6 mM glucose, 0.1% BSA), and stimulated (1 × 10⁶ cells/ml) with either vehicle control or 100 ng/ml dinitrophenyl-human serum albumin (DNP-HSA; 40 mol DNP/mol HSA; Sigma-Aldrich) in Tyrode’s at 37°C for 0–27 min (as indicated in figures). Cells were rinsed in ice-cold PBS supplemented with 100 μM sodium orthovanadate, and then lysed in kinase lysis buffer (KLB; 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1% (v/v) Nonidet P-40, 0.5% (w/v) sodium deoxycholate, 10 μg of aprotinin/ml, 10 μg of leupeptin/ml, 1 mM sodium orthovanadate, 100 μM PMSF). Lysates were clarified by centrifugation to generate soluble cell lysates (SCL). Proteins were resolved by SDS-PAGE and transferred to polyvinylidene difluoride Immobilon P (Millipore) membranes using a Trans-Blot semidry transfer apparatus (Bio-Rad). Following blocking in either 5% (w/w) nonfat milk powder or BSA in Tris-buffered saline/0.1% Tween 20 (TBST), primary Abs were incubated overnight at 4°C. The membranes were then washed in TBST and subsequently probed with the appropriate HRP conjugated donkey anti-rabbit IgG (GE Healthcare) or sheep anti-mouse Ig (GE Healthcare). The membranes were washed again in TBST and then visualized with chemiluminescence reagent (Applied Biological Materials) and x-ray film (Fujifilm).

WT and KO BMMCs were generated following 4OH-TM treatment (100 nM) for 3 days, as described above. Subsequently, BMMCs (5 × 10⁶ cells per sample; 2 × 10⁶ cells/ml) were starved and sensitized with anti-TNP IgE (1 μg/ml; BD Biosciences) for 6 h and then stimulated (2 × 10⁶ cells/ml) with either vehicle control or 10 ng/ml DNP-HSA for 6 h at 37°C. Cells were then pelleted and supernatants were collected for testing SHP2 levels in WT and KO cells. Cell supernatants were analyzed for TNF-α concentration using a commercial ELISA kit (eBioscience), according to the manufacturer’s instructions.

WT and KO BMMCs were generated following 4OH-TM treatment (100 nM) for 3 days, as described above. The BMMCs (10⁵ cells/sample) were then starved and sensitized with anti-TNP IgE (1 μg/ml; BD Biosciences) for 6 h, washed with warm Tyrode’s buffer and then stimulated for 1 h (10⁶ cells/ml) in Tyrode’s buffer supplemented with vehicle control, 10 ng/ml DNP-HSA, or 1 μM calcium ionophore at 37°C. Cells were pelleted, supernatants were collected, and cell pellets were lysed in Tyrode’s buffer supplemented with 0.5% Triton X-100 to generate whole cell lysates. All samples (supernatants and pellet samples) were analyzed for β-hexosiaminidase activity via a colorimetric assay described previously (35).

Densitometry was conducted on autoradiographs with nonsaturated signals, which were scanned and analyzed using histogram function in Corel Photo-Paint (CorelDRAW graphics suite 12, Corel). Relative levels of phosphorylation-specific signal over total protein were calculated, and presented as fold increase in the representative immunoblots. Statistical significance was defined as p ≤ 0.05 using Student’s t test.

To study the function of SHP2 in FcεRI signaling in mast cells, we crossed shp2^fl/fl mice (28) with transgenic mice expressing Cre recombinase-estrogen receptor fusion (CreER)* ubiquitously (32). Bone marrow from shp2^fl/fl and shp2^fl/fl:Tg^CreER* was cultured in the presence of IL-3 and SCF (and absence of 4OH-TM) for 4 wk to generate mature BMMCs. Consistent with tight regulation of the CreER* fusion protein and normal SHP2 function, no evidence of shp2 exon 4 deletion (data not shown), or defects in maturation to FcεRI/Kit-positive BMMCs was observed in the CreER*-positive compared with CreER*-negative BMMCs (Fig. 1,A). 4OH-TM was added to the medium of shp2^fl/fl:Tg^CreER* BMMCs for 3 days to allow shp2 exon 4 deletion and sufficient time for SHP2 protein depletion (hereafter referred to as KO). To control for potential effects of the 4OH-TM treatment, shp2^fl/fl (CreER*-negative) were treated as above (hereafter referred to as WT). At 3 days post 4OH-TM treatment, we observed no difference in surface expression of FcεRI/Kit (Fig. 1,A, lower panels), and an ≈80% decrease in SHP2 protein levels in KO BMMCs, compared with WT BMMCs (Fig. 1 B). No significant differences in cell viability were noted between WT and KO BMMCs between 1 and 4 days post 4OH- TM treatment (data not shown). Thus, although SHP2 regulates myelopoiesis (27) and signaling from receptors for IL-3 (36) and SCF (33), this temporal inactivation approach allows for study of SHP2’s role in FcεRI signaling in mature BMMCs.

To probe the role of SHP2 on FcεRI signaling, WT and KO BMMCs (prepared as above) were starved, sensitized with anti-DNP-IgE, and time courses of DNP-HSA (Ag) treatment were performed. The overall profiles of tyrosine phosphorylated proteins were similar between WT and SHP2 KO BMMCs following FcεRI clustering (Fig. 2, top panel, compare lanes 1–4 and 5–8). To directly test for effects of SHP2 on FcεRI-associated PTKs, we assessed phosphorylation of residues in Lyn and Syk (Y396 and Y346 in mouse, respectively) that correlate with kinase activation (7, 37). In WT BMMCs, we observed a transient increase in pY396-Lyn, which was maximal at 3 min post-FcεRI aggregation (Fig. 2, second panel, lanes 1–4). Interestingly, pY396-Lyn was elevated in magnitude and duration in KO BMMCs compared with WT BMMCs (Fig. 2,B, compare lanes 5–8 with 1–4). Because Syk is a key downstream substrate of Lyn once recruited to FcεRI γ-chain ITAMs, we tested whether Syk phosphorylation was affected by SHP2 depletion. Interestingly, the magnitude and duration of pY346-Syk was significantly increased in SHP2 KO BMMCs at later times, compared with WT BMMCs (Fig. 2, fourth panel). Similar results were obtained with three independent sets of SHP2 WT and KO BMMCs established from three separate pairs of shp2^fl/fl and shp2^fl/fl:Tg^CreER* mice. Taken together, these results suggest that SHP2 is not required for initiation of FcεRI signaling, but may function to limit the duration of Lyn and Syk activation in BMMCs.

Because mast cell degranulation is an early response to FcεRI aggregation, and is highly regulated by tyrosine phosphorylation-based signaling mechanisms, we tested whether SHP2 regulates degranulation in BMMCs. We first tested whether FcεRI-induced phosphorylation of PLCγ, as a Syk substrate and key effector of the calcium-dependent degranulation pathway in mast cells. Although no defects in FcεRI-induced phosphorylation of PLCγ1 were observed (Fig. 3,A, top panel), phosphorylation of PLCγ2 (phosphorylation of pY1217-PLCγ2 correlates with its activation; Ref. 38) was slightly elevated in SHP2 KO BMMCs compared with WT (Fig. 3,A, third panel, compare lanes 1–4 with 5–8). These results suggest that Syk-induced PLCγ activation and the calcium-dependent degranulation pathway does not require SHP2. To measure degranulation SHP2 WT and KO BMMCs were starved, sensitized with anti-TNP IgE, and challenged with DNP-HSA (Ag) or calcium ionophore (A23187) and tested for β-hexosiaminidase release, as previously described (39). No differences in FcεRI-induced, or calcium ionophore-induced β-hexosiaminidase release were observed between SHP2 WT and KO BMMCs (Fig. 3 B). Because SHP2 protein levels were effectively depleted in these cells based on IB (data not shown), we conclude that SHP2 is not a key regulator of FcεRI-induced degranulation.

Members of the p38, Jnk, and Erk MAPK subfamilies play key roles downstream of FcεRI in promoting the de novo production of lipid mediators and numerous cytokines and chemokines (3). To test whether SHP2 regulates FcεRI-induced activation of MAPKs, we surveyed phosphorylation of key activation loop residues in p38, Jnk, and Erk MAPKs in WT and KO BMMCs. We noted no significant differences in the magnitude or kinetics of phosphorylation of p38 between WT and KO BMMCs (Fig. 4, top panels). However, we observed reduced FcεRI-induced phosphorylation of Jnk and Erk MAPKs in SHP2-depleted BMMCs compared with control (Fig. 4, middle and lower panels, compare lanes 1–4 with 5–8). These results are consistent with a positive role for SHP2 promoting Jnk and Erk activation in BMMCs activated via FcεRI. These results were observed in multiple, independent pairs of SHP2 WT and KO BMMCs (data not shown), and we conclude that SHP2 signals upstream of Jnk and Erk MAPKs within the FcεRI pathway, similar to the findings for the SCF/Kit pathway in BMMCs (33).

Because Fyn PTK was shown to positively regulate Erk and Jnk activation in BMMCs (40), and Fyn is negatively regulated by Lyn PTK-mediated phosphorylation of Cbp/PAG and Csk recruitment (41), we hypothesized that Fyn activity may be suppressed in SHP2 KO BMMCs. To test this, we examined phosphorylation of inhibitory tyrosines in Fyn and Lyn (pY531 and pY507, respectively), and phosphorylation of the Fyn substrate Gab2. Although no significant differences were observed in pY507-Lyn between genotypes, a significant increase in pY531-Fyn was observed in SHP2 KO BMMCs compared with WT BMMCs (Fig. 5, compare lanes 1–4 with 5–8). Because phosphorylation of Gab2 adaptor protein has been shown to be dependent on Fyn PTK (42, 43), we tested whether phosphorylation of the PI3K recruitment site (pY441 in mouse; pY452 in human) in Gab2 (36, 44, 45) was altered in SHP2 KO BMMCs following IgE/DNP challenge. Consistent with reduced Fyn activity in SHP2 KO BMMCs, we observed decreased Gab2 phosphorylation in SHP2 KO BMMCs compared with WT BMMCs (Fig. 5, bottom panels, compare lanes 5–8 with 1–4).

Because the Fyn/Gab2/PI3K signaling axis was defective in SHP2 KO BMMCs, and this is expected to limit activation of Akt kinase, we tested whether downstream signaling to Akt is regulated by SHP2 in BMMCs. Interestingly, a significant decrease in phosphorylated Akt (pAkt) was observed in SHP2 KO BMMCs compared with WT (Fig. 6 A, first and second panels). Similar results were obtained with three independent pairs of WT and KO BMMCs, suggesting that SHP2 promotes Akt activation downstream of FcεRI in mast cells. This is likely an indirect effect of impaired Fyn activity in SHP2-depleted BMMCs.

Akt signaling leads to activation of several classes of transcription factors, including NFκB activation (3). NFκB activation depends on activation of IKK, which is an Akt substrate (46), and shown to reside downstream of Fyn PTK (40). Analysis of the phosphorylation of IKK in WT and KO BMMCs was performed and showed that FcεRI-induced IKK phosphorylation was reduced in SHP2 KO BMMCs compared with WT (Fig. 6,A, third and fourth panels, compare lanes 5–8 with 1–4). Next, we examined FcεRI-evoked IκB down-regulation following IKK-mediated phosphorylation. Although IκB levels were slightly higher in SHP2 KO BMMCs compared with WT at resting conditions, the FcεRI-induced down-regulation of IκB was not significantly impaired in KO BMMCs, compared with WT BMMCs (Fig. 6 A, fifth panel; Erk served as a loading control). Thus, although the levels of IKK phosphorylation were reduced in SHP2 KO BMMCs, the remaining activity was sufficient to induce IκB turnover.

Taken together these results suggest that SHP2 promotes activation of Jnk, Erk, and Akt kinases, which all impinge on transcriptional up-regulation of cytokines (e.g., TNF-α) in activated mast cells (3). To investigate this possibility, we compared TNF-α production and release by two independent sets of SHP2 WT and KO (WT1/KO1 and WT2/KO2) BMMCs following IgE/Ag treatment for 6 h. Interestingly, the FcεRI-evoked TNF-α release was significantly reduced in both SHP2 KO1 and KO2 cultures compared with either WT BMMC culture (Fig. 6 B; asterisks indicate a significant difference from WT p < 0.05). Similar results were obtained at later times poststimulation, and with three additional sets of WT and KO cultures (data not shown). These results show that although SHP2 is not essential for FcεRI signaling, SHP2 promotes the production of the inflammatory cytokine TNF-α in BMMCs activated via FcεRI.

Previous studies have described the recruitment of SHP2 phosphatase to a variety of receptors and adaptor proteins in mast cells following Ag-mediated clustering of FcεRI (reviewed in Ref. 3). In this study, we used a transgenic mouse model allowing temporal inactivation of shp2 alleles in mature BMMCs to directly probe the function of SHP2 in FcεRI signaling in BMMCs. Our findings point to a positive role for SHP2 in certain arms of the FcεRI signaling cascade. Consistent with studies of SHP2 in other cell types and pathways (24, 47), we find that SHP2 promotes activation of Jnk and Erk MAPKs, Akt and IKK pathways downstream of FcεRI. This effect is consistent with elevated Lyn activity and reduced Fyn signaling in SHP2 KO BMMCs. Lyn is known to antagonize Fyn PTK via phosphorylation of the transmembrane adaptor Cbp/PAG within lipid rafts (41). Because Cbp/PAG recruits Csk to the membrane allowing for phosphorylation of inhibitory sites in SFKs (48), it will be important to assess the Cbp/PAG/Csk signaling axis in contributing to the phenotypes we observe in SHP2 KO BMMCs. Cbp/PAG was shown to inhibit FcεRI-induced signaling and degranulation in RBL-2H3 cells (49), and has been implicated in the mechanism for Lyn-mediated antagonism of Fyn activity in mast cells (41). In addition, a study of mast cells from epilepsy prone (EL) and resistant (ASK) mice correlated hypophosphorylation of Cbp/PAG with hyperactivation of Fyn in ASK BMMCs, compared with EL BMMCs (50). They also observed prolonged activation of MAPKs and elevated TNF-α secretion in ASK BMMCs compared with EL, suggesting that Fyn functions upstream of these pathways. Another recent study of the contributions of the SFK Hck to mast cell activation via FcεRI, showed that Hck KO BMMCs displayed elevated Lyn activity and hyperphosphorylation of Cbp/PAG (51). In this study, no difference in C-terminal phosphorylation of Lyn was observed in Hck KO BMMCs compared with control cells, consistent with Cbp/PAG/Csk inhibition of other SFKs that promote downstream signaling to MAPKs, Akt, and cytokine production (51). These results are consistent with our observed defects in activation of Fyn, but not Lyn, in SHP2 KO BMMCs. Thus, defects in Fyn or Hck activation in SHP2-depleted BMMCs may explain some of the defects we observe in activation of Jnk and Erk MAPKs, Gab2/Akt/IKK pathway, and TNF-α release. It will be important to test this hypothesis further by direct inhibition of the Cbp/PAG/Csk signaling axis in mast cells.

Another potential target of SHP2 relevant to promoting Ras/Erk signaling based on work in other cell types that could be explored in SHP2-deficient mast cells is RasGAP. SHP2 dephosphorylates residues in receptors and adaptor proteins that allow recruitment of RasGAP to membranes, thus allowing attenuation of Ras signaling (reviewed in Ref. 23). A previous study in RBL-2H3 cells, showed that RasGAP recruitment to Dok family adaptors following FcεRI aggregation, suppresses the Ras/Erk pathway and TNF-α production (52). Although the SH2 domains of SHP2 displayed only weak interactions with Dok-1 (52), it will be important to determine whether Dok adaptors are potential SHP2 substrates using a substrate trapping approach (53). As Ras promotes activation of both the Erk pathway as well as the PI3K/Akt pathway, elevated RasGAP recruitment to the membrane via Dok adaptors in SHP2-deficient BMMCs may contribute to defects in signaling to Jnk, Erk, and Akt pathways that promote cytokine production. Consistent with this hypothesis is the recent finding that SHP2 promotes Ras activation and downstream activation of Rac and Jnk kinase in SCF-treated BMMCs (33).

Mast cells also express SHP1 phosphatase, which although highly related to SHP2 in structure (54) has generally proven to be a negative regulator of signaling from ITAM-containing receptors in B cells, T cells, and mast cells (55). The generally opposing functions of SHP1 and SHP2 is likely explained in part by preferential recruitment of SHP1 to ITIM-containing receptors that serve to dampen ITAM signaling. Differences in SH2 domain specificities between SHP1 and SHP2 account for distinctions between their preferred ligands, subcellular localization, and substrate specificities. However, one recent study in colorectal cancer cells suggests that SHP1 and SHP2 play overlapping functions in promoting Erk activation downstream of EGFR (56). They identified a role for SHP2 in facilitating SHP1 recruitment to phoshorylated Gab1 adaptor via protein-protein interactions between SHP1 and SHP2. However, the domains or residues mediating this interaction were not described. In mast cells, recent studies of SHP1 and SHP2 functions in the FcεRI pathway (Ref. 57) and this study), provide evidence for distinct roles of these related phosphatases. A previous study of motheaten mice that are deficient in SHP1 had implicated SHP1 in repressing allergic inflammation (9). More recently, Nakata and coworkers (57) used BMMCs from motheaten mice to illustrate that SHP1 limits phosphorylation of the Syk substrates SLP-76 and LAT, and suppresses cytokine production. In addition, SHP1 was shown to promote calcium mobilization and FcεRI-evoked degranulation of BMMCs, thus illustrating positive and negative roles in regulating mast cell activation. It is worth noting that in this study, we identified an opposite effect of SHP2 in promoting cytokine production, with no defects in degranulation. Thus, there is currently evidence for unique functions of SHP1 and SHP2 in regulating mast cell activation. However, we cannot exclude the possibility that some overlapping, cooperative functions of SHP1 and SHP2 occur in mast cells. Future studies with compound SHP1/SHP2-mutant mice, or RNA interference approaches in mast cells, will be required to address this possibility directly.

Recently, a transgenic mouse line has been created that drives Cre recombinase expression in mast cells and their progenitors under the control of the Mcpt-5 promoter (58). It will be important in the future to extend our findings on the role of SHP2 in mast cell biology by generating mast cell-specific SHP2 KO mice and study its potential role in mast cell development, homing, migration, and allergic inflammation. However, if Mcpt5-driven Cre expression occurs in early mast cell progenitors, SHP2 inactivation may prevent mast cell differentiation. Also, the key role of SHP2 in integrin signaling (59) and cell migration (60) may adversely affect mast cell homing to target tissues. Nevertheless, the development of a mast cell-specific SHP2 KO mouse model would likely provide valuable new insights into SHP2 function in mast cell-related processes in adult animals.

In conclusion, this study has provided novel insight into the role of SHP2 in regulating FcεRI signaling and mast cell activation. In future, it will be important to extend these experiments to in vivo models and human mast cells. In addition, there is much to learn of the molecular mechanisms involved in the positive regulation of mast cell activation mediated by SHP2, and whether SHP2 is a potential target for new therapeutic approaches to combat allergic inflammation.

We thank Dr. Joan Brugge for providing Lyn and Syk Abs, and members of the laboratory for helpful comments.

The authors have no financial conflict of interest.