SH2 Domain-Containing Phosphatase 2 Is a Critical Regulator of Connective Tissue Mast Cell Survival and Homeostasis in Mice

Abstract

Mast cells require KIT receptor tyrosine kinase signaling for development and survival. Here, we report that SH2 domain-containing phosphatase 2 (SHP2) signaling downstream of KIT is essential for mast cell survival and homeostasis in mice. Using a novel mouse model with shp2 deletion within mature mast cells (MC-shp2 knockout [KO]), we find that SHP2 is required for the homeostasis of connective tissue mast cells. Consistently with the loss of skin mast cells, MC-shp2 KO mice fail to mount a passive late-phase cutaneous anaphylaxis response. To better define the phenotype of shp2-deficient mast cells, we used an inducible shp2 knockout approach in bone marrow-derived mast cells (BMMCs) or cultured peritoneal mast cells and found that SHP2 promotes mast cell survival. We show that SHP2 promotes KIT signaling to extracellular signal-regulated kinase (ERK) mitogen-activated protein kinase and downregulation of the proapoptotic protein Bim in BMMCs. Also, SHP2-deficient BMMCs failed to repopulate mast cells in mast cell-deficient mice. Silencing of Bim partially rescued survival defects in shp2-deficient BMMCs, consistent with the importance of a KIT → SHP2 → Ras/ERK pathway in suppressing Bim and promoting mast cell survival. Thus, SHP2 is a key node in a mast cell survival pathway and a new potential therapeutic target in diseases involving mast cells.

INTRODUCTION

Mast cells (MCs) produce a multitude of mediators during innate and adaptive immune responses that alter vascular permeability, leukocyte recruitment, and activation (17). Although mast cells are key effectors of allergic inflammation, there are numerous reports that mast cells are protective primarily during immune responses to bacteria and parasites and in resolving inflammation in the skin (22). Mast cell progenitors arise from common myeloid progenitors in bone marrow (14), reside in the spleen (5), and traffic to target tissues, such as mucosa and connective tissues, where they differentiate into mature mucosal mast cells (MMCs) and connective tissue mast cells (CTMCs), respectively (21). CTMCs can be distinguished from MMCs on the basis of their mediators stored in granules and their expression profiles of mast cell proteases. In mice, MMCs in the gut express only tryptase family peptidases (e.g., MCP-6, MCP-7), while CTMCs in peritoneum and skin express both tryptase and chymase-like serine peptidases, namely, MCP-4 and MCP-5 (encoded by the Mcpt4 and Mcpt5 genes, respectively) (9).

KIT (CD117) receptor tyrosine kinase signaling is critical for mast cell development and survival of mature mast cells. KIT activation occurs via binding of its ligand stem cell factor (SCF), causing the dimerization and autophosphorylation of KIT (19, 38). Mice with loss-of-function mutations in the white spotting (W) locus, encoding the KIT receptor (e.g., Kit^W-sh/W-sh), or within the steel locus, encoding SCF, exhibit various degrees of mast cell deficiency (7, 20). Further studies using knock-in mutations of individual KIT tyrosine residues identified key signaling pathways required for mast cell development (1, 24, 25, 45, 48). Phosphorylation of KIT Y719 allows recruitment of the p85α regulatory subunit of class I_A phosphatidylinositol 3-kinase (PI3K) (44). Blocking p85α recruitment to KIT Y719 in mice leads to reduced mast cell numbers in peritoneum but not in skin (25). More-severe mast cell deficiencies were reported in p85α null mice than in KIT Y719F mutant mice (15), consistent with an additional PI3K recruitment pathway via the Gab2 adaptor protein downstream of KIT. Yu and coworkers showed that the KIT Y567/Y569-Gab2 and Y719-PI3K pathways have overlapping functions in mast cell development (48). SHP2 (encoded by ptpn11) is recruited to tyrosine-phosphorylated Gab2 following KIT activation and promotes activation of Ras and Rac GTPases, Jun N-terminal protein kinase (Jnk) mitogen-activated protein kinase (MAPK), and proliferation in bone marrow-derived mast cells (BMMCs) (48). However, the role of SHP2 in mast cell development or function in vivo has not been reported.

SHP2 is a nonreceptor protein tyrosine phosphatase (PTP) that signals downstream of diverse receptors in various tissues during embryonic development and in adults (37). In mice, shp2 null mutations result in early embryonic lethality at the peri-implantation stage, with defects in trophoblast stem cell proliferation and survival observed (47). Hypomorphic alleles of SHP2 are also embryonic lethal but die at midgestation (6, 42). More recently, diverse roles of SHP2 in adult mice have been identified by Cre recombinase/LoxP-mediated deletion of shp2 exon 4 or exon 11 (47, 49). In postmitotic neurons, SHP2 regulates energy balance and early-onset obesity (26, 49), whereas in neural stem cells, SHP2 KO results in defective differentiation and early postnatal lethality (23). SHP2 is essential for the survival of hematopoietic stem cells in mice (10, 16). Conversely, expression of human juvenile myelomonocytic leukemia (JMML)-associated shp2 mutants in mouse hematopoietic stem/progenitor cells results in the development of a fatal JMML-like disorder (34, 46). Overexpression of SHP2 mutants leads to hyperactivation of extracellular signal-regulated kinase (ERK) and AKT kinases in IL-3-treated BMMCs (34) and increased survival of hematopoietic stem cells (46). Overall, SHP2 plays key roles in multiple cell types and pathways required for normal physiology and disease pathologies (11, 28).

To directly test the role of SHP2 in mature mast cell homeostasis and function in vivo, we crossed shp2 exon 4 flox mice (49) to a strain with mast cell-specific transgenic expression of Cre, in which Cre is driven by the Mcpt5 promoter (43). Here, we report that CTMC-specific knockout of shp2 (MC-shp2 KO) leads to a severe deficiency in the numbers of mature mast cells in peritoneum and skin, compared to numbers in control mice. Consistent with this defect in mast cell numbers, MC-shp2 KO mice fail to mount an IgE-mediated cutaneous reaction, unlike control mice. To gain more insight into the function of SHP2 in mast cells, we used a temporal deletion approach to generate shp2 KO BMMCs and peritoneal mast cell cultures. SHP2-deficient mast cells showed survival defects upon removal of cytokines. Similar results were observed with SHP2 inhibitor treatment, suggesting that the phosphatase activity of SHP2 contributes to survival signaling in mast cells. SHP2-deficient BMMCs were also defective in the repopulating of mast cells in Kit^W-sh/W-sh mice. Thus, SHP2 promotes the survival of BMMCs in vitro and is critical for CTMC homeostasis in vivo.

MATERIALS AND METHODS

Mice.

Transgenic mice expressing Cre under the control of the Mcpt5 gene regulatory elements (Tg^Mcpt5-Cre) have been described previously (43). Mice with floxed shp2 exon 4 (49) were crossed with Tg^Mcpt5-Cre mice, and their progeny were set up in breeding pairs to generate floxed control (shp2^fl/fl) and experimental (Tg^Mcpt5-Cre:shp2^fl/fl [referred to here as MC-shp2 KO]) mice. Generation of Tg^CreER:shp2^fl/fl and Tg^CreER:shp2^+/+ mice on a C57BL/6 background was described previously (31). Mast cell-deficient Kit^W-sh/W-sh mice on a C57BL/6 background were obtained from The Jackson Laboratory. All animals were housed and maintained at Queen's Animal Care Services. All mouse studies were approved by the Queen's University Animal Care Committee.

Shp2 flox and null allele detection.

The genotyping of shp2 flox and null alleles was described previously (31, 49). For enhanced detection of the shp2 null allele, nested PCR was performed using shp2 null primers for 15 cycles (49), followed by 30 cycles with the following nested shp2 primers: 5′TTCACTAAATGCAACAACTGGC3′ (forward) and 5′GGACAGGTACTAGGCTCCATCCC3′ (reverse).

Derivation of BMMCs.

Femoral bone marrow from shp2^fl/fl, Tg^CreER:shp2^+/+, and Tg^CreER:shp2 ^fl/fl mice were flushed with Iscove's modified Dulbecco's medium (IMDM) complete medium (IMDM, 10% [vol/vol] fetal bovine serum, 1 mM sodium pyruvate [Invitrogen], 1% [vol/vol] antimicrobial-antimycotic solution [Invitrogen], 1% [vol/vol] nonessential amino acids [Invitrogen], 1% [vol/vol] l-glutamine, 2% [vol/vol] conditioned medium from X63 IL-3 cells, and 50 μM α-monothioglycolate [Sigma-Aldrich]) and cultured for 4 to 6 weeks. The maturity of BMMC cultures was determined by a high expression of KIT and FcεRIα via flow cytometry.

Inducible shp2 deletion in BMMCs.

BMMCs were established by culture in IL-3-conditioned medium for 4 to 6 weeks and tested for maturity (KIT⁺ FcεRI⁺), as previously described (31). Mature BMMCs from mice of the shp2^fl/fl, Tg^CreER:shp2^+/+, and Tg^CreER:shp2^fl/fl genotypes were treated with IMDM complete medium supplemented with 200 nM 4-hydroxytamoxifen (4TM; Sigma-Aldrich) for 3 days to derive wild-type (WT) and shp2 knockout (KO) cultures, respectively. For SCF stimulation experiments, BMMCs (5 × 10⁶/sample) were starved of IL-3 for 6 h, washed twice in Tyrode's buffer (10 mM HEPES [pH 7.5], 130 mM NaCl, 5 mM KCl, 1.4 mM CaCl₂,1 mM MgCl₂, 5.6 nM glucose, 0.1% bovine serum albumin), and stimulated or not stimulated with SCF (50 ng/ml; Peprotech) for 5 or 15 min at 37°C. Lysates were prepared and analyzed by immunoblotting as previously described (41).

Immunoblotting.

The following commercial antibodies were employed: rabbit anti-AKT (C67E7, 1:1,000; Cell Signaling Technology), rabbit anti-phospho-AKT (T308, C31E5E, 1:1,000; Cell Signaling Technology), rabbit anti-Bim (C34C5, 1:1,000; Cell Signaling Technology), rabbit anti-ERK (K23; 1:1,000; Santa Cruz Biotech Inc.), rabbit anti-KIT (1:1,000; Cell Signaling Technology), rabbit anti-phospho-KIT (Y719, 1:1,000; Cell Signaling Technology), rabbit anti-SHP2 (C18, 1:1,000; Santa Cruz Biotech Inc.), and mouse anti-α-tubulin (1:1,000; Sigma). Horseradish peroxidase (HRP)-conjugated secondary antibodies to mouse and rabbit IgG were from GE Healthcare, and immunoblots were revealed by enhanced chemiluminescence (ECL) (Perkin Elmer).

Analysis of tissue-resident mast cells in mice.

Thin tissue sections (4 μm) from formalin-fixed, paraffin-embedded skin (ear and back) and glandular stomach were stained with 1% alcian blue (pH 2.5 in 3% acetic acid) and 0.1% nuclear fast red (Kernechtrot; Poly Scientific R&D Corp.). Images were acquired by microscopy, and mast cells were scored for 10 to 15 fields of view for each mouse tissue. Peritoneal cells were obtained by lavage in IMDM complete medium, and peritoneal mast cells were identified by staining them with allophycocyanin (APC)–anti-KIT or fluorescein isothiocyanate (FITC)–anti-KIT and phycoerythrin (PE)–anti-FcεRIα (both from Cedarlane Biotech), followed by fluorescence-activated cell sorting (FACS) analysis. To assess other major cell types, peritoneal cells were also stained with PE–Cy5–anti-B220, FITC–anti-Gr-1, and PE–anti-CD11b (all from Cedarlane Biotech.), followed by FACS analysis. Isotype controls were employed in all experiments.

IgE-mediated late-phase passive cutaneous reaction.

IgE-mediated passive cutaneous reactions, which are dependent on mast cells (8), were performed essentially as described previously (36). Control (shp2^fl/fl) and MC-shp2 KO mice (5 mice per genotype from 2 separate experiments) were sensitized by tail vein injection of antidinitrophenyl (anti-DNP)-IgE (2 μg in 0.1 ml saline). After 24 h, cutaneous reactions were initiated by topical application of 20 μl of dinitrofluorobenzene (DNFB; 0.2% [wt/vol] in acetone-olive oil [4:1]) on the skin of the right ear and the right hind paw of each mouse. As a control, vehicle was applied to the left ear and left hind paw of each mouse. After 24 h, mice were sacrificed and the difference in thickness between DNFB- and vehicle-treated ear skin and paws was measured using digital calipers. Tissues were formalin fixed and paraffin embedded, and sections were stained with hematoxylin and eosin (H&E). Images were acquired with an Olympus BX51 microscope equipped with a Q Color5 digital camera (20× objective; images were acquired using QCapturePro software).

Repopulating mast cells in Kit^W-sh/W-sh mice.

Mature BMMCs derived from shp2^fl/fl and Tg^CreER:shp2^fl/fl mice were treated with 200 nM 4TM for 3 days to generate WT and shp2 KO cultures, respectively. Kit^W-sh/W-sh mice (6 per genotype; 4 to 6 weeks old) were reconstituted with WT and shp2 KO BMMCs by intraperitoneal (i.p.) (5 × 10⁶ BMMCs in 200 μl phosphate-buffered saline [PBS]) and intradermal (i.d.) injections in ears (5 × 10⁵ BMMCs in 20 μl PBS). Recipient mice were analyzed after day 1 and after 4 weeks for mast cell numbers in peritoneum and ear skin as described above (3 experimental replicates, with similar results observed).

BMMC survival assays.

For survival assays, WT and KO BMMCs (5 × 10⁵/sample; prepared as described above) were maintained in complete BMMC medium (IL-3), without IL-3 or SCF (starved), or with SCF (25 or 50 ng/ml) for 72 h. Cell survival was assessed by staining cells with FITC-annexin V (BD Biosciences) and propidium iodide, followed by FACS analysis. Survival assays were also conducted to compare the effects of vehicle (dimethyl sulfoxide [DMSO]) or SHP2 inhibitor II-B08 (50) using WT BMMCs. Following cytokine starvation overnight, BMMCs were treated with vehicle or II-B08 (15 μM; added fresh each day) for 72 h, prior to FITC-annexin V or propidium iodide staining and FACS analysis as described above.

BMMC lentiviral transduction and survival assays.

Lentiviral vectors (pGIPZ) encoding a green fluorescent protein (GFP) reporter cassette along with short hairpin RNAs (shRNAs) against mouse Bim catalog number RMM 3532; Open Biosystems) or a nontargeting (NT) control (clone identifier, RHS 4346; Open Biosystems) were obtained. Lentiviruses were produced by transfection of HEK293 cells (grown on 100-mm plates) with a pGIPZ-based plasmid (4.2 μg), pCMVΔR8.91 packaging plasmid (4.2 μg), and pMD.2G envelope plasmid (2 μg) using TransIT-LT1 transfection reagent (Mirus). Two batches of conditioned medium were collected at 48 and 72 h, filtered through 0.45-μm sterile filters, and stored at −80°C in aliquots. Virus titer was determined by transduction of P815 mouse mastocytoma cells followed by FACS analysis (percent GFP positive). Following selection with puromycin (2 μg/ml) for 5 days, cell pools were derived for each virus and lysates obtained. Immunoblots with anti-Bim (C34C5; Cell Signaling Technology) were performed, and the most-effective Bim shRNAs were identified as shRNA1 (clone B3, target sequence, 5′CGGCCTTTGTACTTAAATATAT3′; V2LMM_220682), shRNA2 (clone G8, target sequence, 5′AGCGGATCGGAGACGAGTTCAA3′, V3LMM_483928). For studies of BMMCs, lentiviral transduction of BMMCs from Tg^CreER:shp2^fl/fl mice was achieved by spin infection (2 × 10⁶ cells/viral supernatant sample; 1,500 × g for 90 min). BMMCs were treated with or without 4TM to generate KO or WT cultures, respectively, and subjected to survival assays as described above (except with PE-conjugated annexin V).

PCMC cultures.

Peritoneal-cell-derived mast cell (PCMC) cultures were established as described previously (30). Briefly, peritoneal cells from shp2^fl/fl, Tg^Mcpt−Cre:shp2^fl/fl, Tg^CreER:shp2^+/+, and Tg^CreER:shp2^fl/fl mice were collected by lavage with IMDM complete medium supplemented with 5%-SCF-conditioned medium (vol/vol) obtained from HEK293-KLS cells. PCMC cultures were supplemented with fresh medium after every 4 days until mature cultures (4 weeks) were obtained. The maturity of PCMC cultures was determined by high expression of KIT and FcεRIα via flow cytometry.

PCMC viability assay.

alamarBlue was used to assess viability or survival, based on color change in a reducing environment of metabolically active cells (3). Assays were conducted per the manufacturer's instructions (Invitrogen). Briefly, PCMCs (8 × 10⁴ cells/well) with or without cytokines were grown for 72 h in 100 μl of IMDM complete medium devoid of IL-3 and SCF in a 96-well flat-bottom plate. alamarBlue (10 μl) was added to each well for 12 h, prior to the reading of absorbance at 570 nm (reference wavelength, 600 nm) using a plate-reading spectrophotometer (Multiskan Spectrum; Thermo Electron Corp.). Reference corrected values were subtracted from the value for a blank (DMEM plus alamarBlue) and are expressed as percentages of reduced (viable) cells relative to values for shp2^fl/fl PCMCs in cytokine-containing medium.

Statistical analysis.

Differences between experimental groups were tested for statistical significance (defined as a P of ≤0.05 [indicated by * in the figures] or a P of ≤0.01 [indicated by **]) using a paired Student t test in Excel (Microsoft, Redmond, WA).

RESULTS

SHP2 is required for peritoneal mast cell homeostasis.

To investigate the role of SHP2 in mature CTMCs in vivo, we crossed Mcpt5-Cre transgenic (Tg^Mcpt5-Cre) mice (43) with shp2^fl/fl mice (49) to generate mast cell-specific shp2 KO (MC-shp2 KO) mice. MC-shp2 KO mice are viable with no overt phenotypes, consistent with the restricted expression of Cre in CTMCs that was previously described (13, 43). To test for evidence of shp2 deletion, we isolated genomic DNA from bone marrow (BM), peritoneal cells (PCs), and several other tissues from Tg^Mcpt5-Cre mice with wild-type (+/+), heterozygous (+/fl), or homozygous (fl/fl) shp2 alleles. Using the previously published method (49), we detected the expected pattern of shp2 WT and flox alleles in PCR samples from BM cells and PCs (Fig. 1A, lower panel). As expected, shp2 null alleles were detected in shp2^+/fl and shp2^fl/fl PC samples, along with samples from a positive control prepared from 4-hydroxytamixofen (4TM)-treated Tg^CreER:shp2^fl/fl BMMCs (Fig. 1A, top panel). We did not detect shp2 null alleles in BM cells (Fig. 1A), which is consistent with Mcpt5-Cre-driven expression of an enhanced yellow fluorescent protein (EYFP) reporter in mature CTMCs but not in BM cell progenitors or MMCs (13, 43). To confirm the specificity of shp2 deletion, we characterized shp2 alleles in PCs from heterozygous mice (with or without the Mcpt5-Cre transgene), which were sorted into negative and positive pools based on peritoneal mast cell markers (KIT⁺/FcεRIα⁺ cells). shp2 null alleles were strongly detected in peritoneal mast cells and undetectable in other peritoneal cells or in peritoneal cells from mice lacking the Mcpt5-Cre transgene (Fig. 1B). As reported previously (43), no loss of peritoneal mast cells is associated with expression of the Mcpt5-Cre transgene (Fig. 1B).

Fig 1.

SHP2 is required for the maintenance of peritoneal mast cells in vivo. (A) shp2 alleles were assessed by PCR analysis of genomic DNA isolated from bone marrow and peritoneal lavage fluid from wild-type (Tg^Mcpt5-Cre:shp2^+/+), heterozygous (Tg^Mcpt5-Cre:shp2^+/fl), and homozygous (Tg^Mcpt5-Cre:shp2^fl/fl or MC-shp2 KO) mice. All bands corresponded to the expected sizes of PCR products for the shp2 null allele (1 kb), the wild-type allele (0.3 kb), and shp2 flox alleles (0.4 kb) (see the arrows on the right). Genomic DNA from 4TM-treated BMMCs (Tg^CreER:shp2^fl/fl) was included as a positive control (pos ctl). (B) Peritoneal mast cells (KIT⁺/FcεRIα⁺ cells) in peritoneal lavage fluid from shp2^+/fl or Tg^Mcpt5-Cre:shp2^+/fl mice were sorted by FACS and shp2 alleles analyzed as described above for positive and negative cells. (C) Representative FACS histograms for peritoneal mast cells (KIT⁺/FcεRIα⁺ cells) from control (shp2^fl/fl) and MC-shp2 KO (Tg^Mcpt5-Cre:shp2^fl/fl) mice (percentages of total cells). (D) Absolute numbers of peritoneal mast cells (means ± standard errors of the means [SEM]) were scored for 6 mice of each genotype. A significant difference between genotypes was observed (P < 0.01; indicated by asterisks).

To test the potential effect of shp2 deletion on peritoneal mast cells, we compared peritoneal mast cell populations in control mice (shp2^fl/fl) and MC-shp2 KO mice by FACS and observed a striking defect in mast cells in MC-shp2 KO mice (Fig. 1C). This defect in mast cells correlates with shp2 genotype and is not due to Mcpt5-Cre transgene expression (see Fig. S1A in the supplemental material). No differences in other resident peritoneal cell types, including peritoneal macrophages, neutrophils, and B cells, were observed in MC-shp2 KO mice (see Fig. S1B). Quantification of these results for multiple mice of each genotype revealed a significant decrease in absolute numbers of peritoneal mast cells upon SHP2 deletion (Fig. 1D) (6 mice/genotype). This is not a complete loss of peritoneal mast cells in MC-shp2 KO mice, as a limited number of PCMCs were generated from these mice (see Fig. S2A in the supplemental material). However, these cells showed only partial deletion of shp2, evidenced by the remaining flox alleles in these SCF-driven cultures (see Fig. S2B). Taken together, these results indicate that SHP2 is required for peritoneal mast cell homeostasis.

Reduced mast cells in skin but not mucosa of MC-shp2 KO mice.

To further characterize the MC-shp2 KO model, shp2 deletion was tested in other tissues. We detected shp2 null alleles in genomic DNA from lung, ear, and back skin tissues but not in liver from heterozygous or homozygous mice (Fig. 2A). This is consistent with restricted Mcpt5-Cre-driven shp2 deletion in CTMCs (13). Next, we compared the numbers of mast cells in skin sections stained with alcian blue for control and MC-shp2 KO mice. Interestingly, MC-shp2 KO mice had fewer mast cells in ear skin than controls (Fig. 2B, see arrows). Quantification of these results for multiple mice revealed a significant reduction in mast cells (≈3-fold) in the ears of MC-shp2 KO mice compared to those of control mice (Fig. 2C) (9 mice/genotype). Mast cell defects in back skin tissue sections from MC-shp2 KO mice were similar to those of controls (Fig. 2D, see arrows). Quantification of these results revealed a significant reduction in the average numbers of mast cells per field within the dermis and hypodermis of MC-shp2 KO mice (Fig. 2E and F) (6 mice/genotype). Taken together, these results are consistent with SHP2 playing a key role in KIT signaling, which is required for mast cell homeostasis within connective tissues.

Fig 2.

SHP2 is required for mast cell homeostasis in ear and back skin. (A) Genomic DNA was isolated from lung, liver, ear, and back skin tissue samples from control (shp2^fl/fl) and MC-shp2 KO (Tg^Mcpt5-Cre:shp2^fl/fl) mice and analyzed for shp2 null and flox alleles by nested PCR. A no-template control (neg ctl) and a positive control (pos ctl) from 4TM-treated Tg^CreER:shp2^fl/fl BMMCs were also included. The positions of shp2 null, flox, and wild-type (wt) alleles are indicated by arrows on the right. The positions of DNA size markers are shown on the left in kilobases. (B) Ear tissue sections from shp2^fl/fl and MC-shp2 KO mice were stained with alcian blue-nuclear fast red to visualize mast cells (indicated by arrows). (C) Dot plot representing the average numbers of mast cells per field (10 to 15 fields/mouse; 9 mice per genotype), which were significantly reduced in MC-shp2 KO mice compared to in control mice (P < 0.01). (D) Mast cell staining in back skin tissue sections from shp2^fl/fl and MC-shp2 KO mice (indicated by the arrows). (E and F) Dot plots representing the average numbers of mast cells per field (10 to 15 fields; 6 mice per genotype) in dermis (E) and hypodermis (F). The horizontal lines in the dot plots represent the means, which were significantly reduced in MC-shp2 KO mice compared to in control mice (P < 0.01).

To further characterize the mast cell populations in MC-shp2 KO mice, we scored mast cells in the glandular stomachs of control and MC-shp2 KO mice (Fig. 3A). In the mucosa of MC-shp2 KO mice, the numbers of mast cells that we detected were similar to those in control mice (Fig. 3B), but in the submucosa in MC-shp2 KO mice, there were reduced numbers of mast cells (Fig. 3C). This is consistent with shp2 deletion in CTMCs and not in MMCs. This is consistent with Mcpt5-Cre transgenic mice driving Cre expression in CTMCs and not in MMCs (13). Thus, the deficit in mast cells in the connective tissues of MC-shp2 KO mice is likely due to growth or survival defects within the mature CTMCs and not due to defects in mast cell progenitors that retain SHP2 function and allow normal MMC development.

Fig 3.

MC-shp2 KO mice have normal numbers of mucosal mast cells in stomach. (A) Glandular stomach tissue sections of shp2^fl/fl and MC-shp2 KO mice were stained with alcian blue-nuclear fast red to visualize mast cells (CTMCs are indicated by open arrows, and MMCs are indicated by black arrows). (B) The numbers of mast cells per field (means ± standard deviations [SD]; 10 fields/mouse; 3 mice per genotype) were scored in mucosa and submucosa. A significant difference between genotypes (P < 0.01; indicated by the asterisk) was observed in the submucosa.

MC-shp2 KO mice fail to mount an IgE-mediated late-phase cutaneous response.

To directly address whether MC-shp2 KO mice have a functional defect in CTMCs, we analyzed their response to IgE-mediated late-phase reactions in skin, which is an IgE- and mast cell-dependent response (8, 13). Control (shp2^fl/fl) and MC-shp2 KO mice were sensitized with anti-DNP–IgE prior to the induction of cutaneous reactions with topical application of dinitrofluorobenzene (DNFB) (right ear and paw) or the vehicle control (left ear and paw). After 24 h, we detected swelling of DNFB-treated ears and paws of control mice but not of control vehicle-treated tissues (Fig. 4A). In contrast, MC-shp2 KO mice showed a significant defect in the swelling of the DNFB-treated ears and paws, compared to that of control mice (Fig. 4A) (5 mice/genotype). This defect was further evident in tissues sections, in which DNFB-treated skin showed extensive inflammatory infiltrates and edema compared to vehicle-treated skin (Fig. 4B). In contrast, no overt changes in tissue swelling or histology were observed for DNFB-treated skin from MC-shp2 KO mice (Fig. 4B). The severity of the defective cutaneous reactions in MC-shp2 KO mice were similar to those recently reported for new models of inducible and constitutive mast cell deficiency (13). Taken together, these results are consistent with reduced numbers of skin mast cells in MC-shp2 KO mice and point to a functional defect in mast cell-mediated responses in these mice.

Fig 4.

MC-shp2 KO mice are defective in the IgE-mediated late-phase cutaneous response. Control (shp2^fl/fl) and MC-shp2 KO mice (5 mice/genotype, from 2 experiments) were sensitized by tail vein injection with 2 μg anti-DNP–IgE 24 h prior to topical application of DNFB (0.2% [wt/vol]) to the right ear and paw or vehicle (acetone-olive oil, 4:1) to the left ear and paw. After 24 h, mice were culled, and differences in the thicknesses of DNFB-treated and vehicle-treated ears and paws were recorded with digital calipers. (A) The average increases in thickness are presented (±SD) for ears and paws. A significant difference was observed between genotypes (* indicates a P of <0.05; ** indicates a P of <0.01). (B) Histologic changes in the ears of control and MC-shp2 KO mice treated as described above. Representative tissue sections stained with H&E are shown. Scale bar, 100 μm.

SHP2 promotes mast cell survival signaling on the SCF/KIT axis.

To better characterize the phenotype of SHP2-deficient mast cells, we used 4TM treatment to induce shp2 deletion (KO) in BMMCs from Tg^CreER:shp2^fl/fl mice and compared to them to 4TM-treated control BMMCs from shp2^fl/fl mice (WT). As we reported previously (31), no defects in KIT or FcεRI expression levels were observed in shp2 KO BMMCs (data not shown). In these cultures, shp2 KO BMMCs display almost complete conversion of shp2 flox alleles to null alleles upon 4TM treatment (Fig. 5A). This leads to a significant reduction in SHP2 protein levels in shp2 KO BMMCs compared to WT BMMCs (Fig. 5B). To further characterize the role of SHP2 in KIT signaling, we analyzed SCF-induced KIT autophosphorylation at Y719 (pY719) and observed no defects between genotypes (Fig. 5C). KIT signaling to AKT kinase was slightly reduced at early time points in shp2 KO BMMCs but not at later times (Fig. 5C). However, SCF/KIT signaling to ERK kinase was significantly reduced in shp2 KO BMMCs compared to that in WT BMMCs (Fig. 5C).

Fig 5.

SHP2 promotes KIT signaling to ERK kinase to suppress Bim in mast cells. BMMCs from shp2^fl/fl and Tg^CreER:shp2^fl/fl mice were treated with 4TM (200 nM) for 3 days to generate WT and shp2 KO BMMCs, respectively. (A) PCR analysis of shp2 null and flox alleles in genomic DNA isolated from WT and shp2 KO BMMCs. (B) Lysates were prepared from WT and KO BMMCs and subjected to immunoblotting (IB) with SHP2 and ERK antibodies. Positions of relative mass markers (in kilodaltons) are shown on the left. (C) BMMCs were starved of IL-3 for 6 h and treated or not treated with SCF (50 ng/ml) for 5 and 15 min. Lysates were prepared and subjected to IB with phospho-Y719-KIT (pY719-KIT), KIT, phospho-ERK (p-ERK), ERK, phospho-AKT (pT308-AKT), AKT, Bim, and tubulin antibodies. Densitometry was performed to determine the relative phosphorylation levels (the WT level at 5 min was set as 1.0; Bim levels are relative to tubulin levels). Results are representative of three separate experiments. Positions of relative mass markers (in kilodaltons) are shown on the left.

In trophoblasts, SHP2 promotes survival by enhancing ERK activation and downregulating the proapoptotic protein Bim (47). In mast cells, Bim plays a key role in triggering apoptosis following cytokine withdrawal and is downregulated by SCF signaling to the ERK and AKT pathways (4, 35). Next, we tested whether SHP2 regulates Bim levels in mast cells. Interestingly, levels of Bim isoforms (Bim_EL, Bim_L, and Bim_S [alternative splice variants encoding 196, 140, and 110 amino acids, respectively]) were significantly elevated in shp2 KO BMMCs compared to WT BMMCs (Fig. 5C). These results suggest that SHP2 promotes SCF/KIT signaling to ERK kinase and suppression of proapoptotic Bim protein in mast cells.

To test the relevance of Bim upregulation in shp2 KO BMMCs to their phenotype, we assayed mast cell survival in the absence and presence of the cytokines IL-3 and SCF. WT (Tg^CreER:shp2^+/+) and shp2 KO BMMCs were cultured for 72 h in the absence of cytokines (starved cells) or in the presence of SCF (25 or 50 ng/ml) or maintained in complete medium (IL-3). Interestingly, shp2 deletion resulted in a significant increase in cell death (percentage of cells that were annexin V positive), compared to that of WT BMMCs (Fig. 6A). Similar results were obtained with shp2^fl/fl BMMCs as control cells (data not shown), suggesting that shp2 deletion, and not CreER expression alone, causes the survival defects in this model. The prosurvival signals transduced by SCF/KIT and IL-3 or IL-3 receptors were partially impaired in shp2 KO BMMCs compared to control BMMCs (Fig. 6A). We prepared cell lysates from BMMCs that were treated as described above for the survival assay but harvested at 24 h (compared to 72 h in Fig. 6A). We observed lower levels of Bim in BMMCs growing in IL-3 than in those starved of cytokines (Fig. 6B). Interestingly, Bim levels were elevated in shp2 KO BMMCs compared to levels in control cells, with the largest differences in starved cells (in Fig. 6B, tubulin served as a loading control). Thus, SHP2 enhances ERK activation and survival, at least in part, by promoting Bim downregulation in cultured mast cells.

Fig 6.

SHP2 suppresses Bim and promotes the survival of BMMCs. (A) WT (Tg^CreER:shp2^+/+) and shp2 KO BMMCs (prepared as described above) were starved of cytokines, treated with SCF (25 ng/ml or 50 ng/ml), or maintained without cytokine starvation (IL-3) for 72 h. Cell survival was assayed by FITC-annexin V and propidium iodide staining and FACS analyses in triplicate. The graph depicts percentages of annexin V⁺ cells (means ± SD) under each culture condition. A significant difference between genotypes was observed (P < 0.01; indicated by an asterisk). (B) WT and shp2 KO BMMCs (prepared as described above) were starved of cytokines, treated with SCF (25 ng/ml or 50 ng/ml), or maintained without cytokine starvation (IL-3) for 24 h. Lysates were prepared and subjected to IB with Bim and tubulin antibodies. Positions of relative mass markers (in kilodaltons) are shown on the left. (C) shp2^fl/fl BMMCs were starved of IL-3 overnight and maintained in either starvation medium (starved), SCF-containing medium (50 ng/ml [SCF-50]), or complete medium (IL-3) for 72 h. Cells were treated with either the vehicle (DMSO) or II-B08 (15 μM) for 3 days (added each day). Cell survival was assayed by FITC-annexin V-propidium iodide staining and FACS analyses in triplicate. The graph depicts percentages of annexin V⁺ cells (means ± SD) under each culture condition. A significant difference between treatments was observed for starved cells (P < 0.05; indicated by an asterisk).

With the recent development of the cell-permeable SHP2 inhibitor II-B08 (50), we tested whether SHP2 regulates mast cell survival via its phosphatase activity. WT BMMCs were subjected to survival assays as described above, with daily addition of vehicle (DMSO) or II-B08 (15 μM), and annexin V-positive cells were analyzed by flow cytometry. Interestingly, SHP2 inhibitor caused increased apoptosis of BMMCs starved of cytokines but had no effect under SCF or IL-3 growth conditions (Fig. 6C). These results suggest that SHP2 regulates mast cell survival via dephosphorylation of substrates that are important for survival signaling.

To determine whether SHP2 also regulates survival in CTMCs, we established SCF-driven PCMC cultures from control (shp2^fl/fl) and Tg^CreER:shp2^fl/fl mice prior to 4TM treatment (WT and KO, respectively). Genomic PCR of PCMCs revealed that 4TM treatment of KO cultures produced shp2 null alleles and a partial reduction in flox alleles (Fig. 7A). PCMC cultures from both genotypes yielded mature mast cells (Fig. 7B). Consistently with our studies of BMMCs (Fig. 6), we observed defects in the viability of KO PCMCs in both the presence and the absence of SCF (Fig. 7C). Since CTMCs are highly dependent on the SCF/KIT signaling axis for survival, these results are consistent with mast cell deficiency in MC-shp2 KO mice.

Fig 7.

SHP2 promotes the survival of cultured peritoneal mast cells. (A) Peritoneal cells from shp2^fl/fl and Tg^CreER:shp2^fl/fl mice were cultured in SCF to generate PCMCs and treated with 4TM to generate WT and KO mice, respectively. Genomic DNA was analyzed by PCR for shp2 null and flox alleles. (B) The surface expression of KIT and FcεRI on PCMCs of both genotypes was measured by flow cytometry. (C) The viability of WT and KO PCMCs was assessed with or without SCF (50 ng/ml) using alamarBlue, as described in Materials and Methods (* indicates a significant difference between genotypes, P < 0.05).

SHP2 is required for repopulating CTMCs in mast cell-deficient mice.

To determine whether these defects in the survival of shp2 KO mast cells in vitro coincides with mast cell homeostasis defects in vivo, we attempted to repopulate Kit^W-sh/W-sh mice with WT and shp2 KO BMMCs by intraperitoneal and intradermal injections. While Kit^W-sh/W-sh mice injected with WT and KO BMMCs showed a detectable population of mast cells 1 day after injection (Fig. 8A, upper panels), no such population was observed for shp2 KO BMMCs at 30 days (Fig. 8A, lower panels). Quantification of these results from two separate experiments revealed a significant defect in the ability of shp2 KO BMMCs to repopulate peritoneal mast cells, compared to that of WT BMMCs (Fig. 8B) (8 mice per BMMC genotype). To further test the requirement of SHP2 for repopulating mast cells in skin, we also intradermally injected Kit^W-sh/W-sh mice with WT and shp2 KO BMMCs. After 4 weeks, alcian blue-stained mast cells were readily detectable in ear skin injected with WT BMMCs (Fig. 8C). In contrast, practically no skin mast cells were observed for mice injected with shp2 KO BMMCs (Fig. 8C). Quantification of these results from two separate experiments revealed a significant defect in the ability of shp2 KO BMMCs to repopulate skin mast cells, compared to that of WT BMMCs (Fig. 8D). Taken together, these findings provide novel evidence that SHP2 plays an essential role in promoting CTMC survival and homeostasis in vivo.

Fig 8.

SHP2-deficient mast cells fail to repopulate mast cell-deficient mice. (A) Kit^W-sh/W-sh mice were injected (intraperitoneally [i.p.]) with WT and shp2 KO BMMCs (6 mice per BMMC genotype from 2 experimental replicates), as described in Materials and Methods. Representative FACS histograms for peritoneal mast cells (KIT⁺/FcεRIα⁺ cells) in peritoneal lavage fluid from Kit^W-sh/W-sh mice 4 weeks after i.p. injection with WT or shp2 KO BMMCs. (B) The percentage of mast cells (mean ± SEM) within the peritoneum was scored (percentage of total cells; n = 8). A significant difference between BMMC genotypes was observed (P < 0.01; indicated by asterisks). (C) Representative images of alcian blue-nuclear fast red staining of ear skin from Kit^W-sh/W-sh mice 4 weeks after intradermal (i.d.) injection with WT or shp2 KO BMMCs. Arrows indicate positions of mast cells. Scale bar, 100 μm. (D) The graph depicts numbers of mast cells per field (means ± SEM) detected in the ear skin (5 fields/mouse; n = 6) 4 weeks after injection (i.d.). A significant difference between genotypes was observed (P < 0.01; indicated by asterisks).

Bim silencing can partially rescue survival defects in SHP2-deficient mast cells.

To directly test whether SHP2 suppression of Bim in mast cells can enhance their survival, we transduced mature BMMCs from Tg^CreER:shp2^fl/fl mice with lentiviruses encoding a GFP reporter and shRNAs to Bim or a nontargeting (NT) control. Untreated BMMCs served as WT controls, and 4TM treatment generated shp2 KO cells (Fig. 9A). Interestingly, we detected a slight reduction in the number of annexin V/GFP-positive shp2 KO BMMCs with Bim shRNA virus compared to levels in an NT control under both starvation and SCF conditions (Fig. 9A). Quantification of the annexin V median fluorescence intensity (MFI) for GFP⁺ cells revealed a significant increase in the binding of annexin V to KO cells compared to that to WT cells transduced with an NT control (Fig. 9B). Bim silencing in KO BMMCs (with two separate shRNAs) corrected this defect in WT BMMC levels (Fig. 9B). Using P815 mouse mastocytoma cells transduced with this panel of lentiviruses, we detected reduced Bim levels with Bim shRNA1 and shRNA2 compared to those of an NT control (Fig. 9C). Taken together, our results indicate that SHP2 is a key amplifier of SCF/KIT signaling to the Ras/ERK pathway, which inhibits Bim from promoting the survival of mast cells (Fig. 9D).

Fig 9.

Bim silencing partially rescues survival defects in shp2 KO BMMCs. (A) BMMCs from Tg^CreER:shp2^fl/fl mice were transduced with GFP-expressing lentiviruses to drive the expression of nontargeting (NT) or Bim shRNAs, which were not treated or treated with 4TM (200 nM for 3 days) to generate WT or KO cells, respectively. Following removal from IL-3, cells were incubated in either starvation medium or SCF (50 ng/ml) for a further 72 h. Representative histograms for PE-annexin V binding and GFP fluorescence are shown. (B) The graph depicts annexin V MFIs for GFP⁺ cells (means ± SD) for triplicate samples of the indicated genotype and shRNA. (C) Immunoblot (IB) analysis of Bim levels in P815 cells stably transduced with the indicated lentiviruses (ERK IB served as a loading control). Positions of relative mass markers (in kilodaltons) are shown on the left. (D) Simplified pathway model relating SHP2 to mast cell survival. SHP2 is a positive regulator of SCF/KIT signaling to the Ras/ERK pathway, which suppresses the proapoptotic Bim protein in mast cells.

DISCUSSION

KIT receptor tyrosine kinase signaling is critical for mast cell development and survival (32). Since mast cells play key roles in both mounting and controlling inflammation (22), it is important to define key nodes in the KIT signaling pathway that may be exploited to either elevate or dampen mast cell numbers or mediator release. Previously, SHP2 phosphatase was implicated in promoting KIT signaling to the Rac/JNK pathway, which drives proliferation of BMMCs (48). Here, we provide new insights into how SHP2 contributes to KIT signals required for mast cell homeostasis in connective tissues. Using the Mcpt5-Cre system (13, 43) and shp2 flox mice (49), we generated mast cell-specific shp2 KO mice. These mice have severe defects in mature CTMCs in skin and peritoneum compared to littermate controls. Mast cells are known to be required for IgE-mediated late-phase cutaneous responses in mice (8), and consistently with our scoring of skin mast cells in MC-shp2 KO mice, these mice fail to mount a late-phase passive cutaneous anaphylaxis reaction. From these observations, we conclude that SHP2 is a key downstream effector of KIT signaling to maintain the CTMC population in mice. To gain insight into the molecular mechanisms, we used an inducible shp2 KO approach in BMMCs and PCMCs. These studies revealed that SHP2 promotes the survival of mast cells in vitro, at least in part by suppressing Bim levels. Bim was previously implicated in the apoptosis of mast cells following cytokine withdrawal, and this was countered by SCF signaling to MEK/ERK and PI3K/AKT pathways (4, 35). Since SHP2-deficient BMMCs fail to repopulate mast cells in Kit^W-sh/W-sh mice, we conclude that SHP2 is required for mast cell survival in vivo. Direct testing of our hypothesized KIT → SHP2 → ERK → Bim pathway (Fig. 9D) using lentiviral delivery of Bim shRNAs in shp2 KO BMMCs revealed that Bim silencing can reduce annexin V binding to WT levels. Taken together, our results identify SHP2 as a key node in the SCF/KIT signaling axis, governing mast cell survival and homeostasis.

Previous studies using knock-in approaches have identified KIT signaling pathways from Y567 and, to a lesser extent, Y719 as critical for the development of peritoneal and dermal mast cells (1, 24, 25, 45, 48). Gab2 adaptor protein signals downstream of Y567 and SFKs and is essential for SHP2 recruitment to the KIT pathway (48). PI3K is recruited directly to KIT Y719 (44) and also via binding to Gab2. Both of these recruitment sites must be lost to result in a mast cell defect similar to that observed in p85α^−/− class I_A PI3K-deficient mice (25, 48). The MC-shp2 KO mice have a deficit in peritoneal mast cells similar to that described for mice lacking KIT Y567 signaling (1). This is consistent with SHP2 being a critical downstream effector of the KIT Y567 pathway. The remaining dermal mast cells in MC-shp2 KO mice likely reflect that the KIT Y719/PI3K pathway remains intact. Consistently, we observed no change in KIT pY719 levels or downstream activation of AKT after SCF stimulation in SHP2-deficient BMMCs (this paper). A previous study of SCF/KIT signaling in SHP2-deficient BMMCs reported defects in signaling to Ras, Rac, and JNK and reduced proliferation (48). The current study provides evidence that SHP2 deletion in BMMCs also leads to survival defects, likely due to reduced targeting of Bim to the proteosome following phosphorylation by ERK or JNK kinases in SHP2-deficient mast cells, as shown in trophoblast stem cells (47). Bim phosphorylation by ERK or JNK triggers its degradation, leading to increased survival signaling by Bcl-2/Bcl-X_L proteins (29). Bim downregulation via the SCF/KIT axis is important for survival in normal and malignant mast cells (2, 35). Indeed, Bim levels were elevated in shp2 KO BMMCs, compared to in control BMMCs, and this correlated with reduced survival during cytokine withdrawal (this paper). The survival defect in shp2 KO BMMCs was only partially rescued by SCF or IL-3, which suggests that SHP2 promotes survival signaling downstream of both KIT and IL-3 receptors. This role likely involves SHP2 phosphatase activity, since we observed reduced survival of BMMCs in the absence of IL-3 or SCF (this paper). Further studies will be required to identify the SHP2 substrates that are involved in prosurvival signaling. Consequently, mast cells lacking SHP2 failed to repopulate mast cell-deficient mice, suggesting that SHP2 is critical for the survival of mast cells in vivo. Although elevated Bim levels may largely explain this defect, it is unlikely to be the sole contributor, since Bim silencing led to only a partial rescue of trophoblast survival (47). Indeed, we show that Bim silencing in shp2 KO BMMCs could reduce annexin V binding to levels of WT cells but could not protect all cells from apoptosis. Interestingly, a recent study implicates enhanced proteosomal degradation of Bim as a key factor in promoting the survival of mast cell leukemia cells with oncogenic KIT^D816V alleles (2). Thus, SHP2 may also be a valid therapeutic target to limit the survival of malignant mast cells.

SHP2 may also regulate mitochondrial functions directly, since SHP2 protein and activity were detected in mitochondrial fractions (40). In fibroblasts, gain or loss of SHP2 function resulted in defects in cytochrome c oxidase activity, reactive oxygen species, and mitochondrial membrane potential (27). Src kinase also localizes to mitochondria and regulates cytochrome c oxidase activity in osteoclasts (33, 39). SHP2 can promote the activation of Src family kinases via dephosphorylating inhibitory C-terminal phosphorylation sites directly (16) and indirectly by blocking recruitment of C-terminal Src kinase (12). Whether SHP2 regulates the phosphorylation of mitochondrial proteins directly in mast cells will require further studies.

Here we describe the utility of Mcpt5-Cre transgenic mice (43) to generate a novel CTMC-restricted KO model that provides new insights into mast cell biology. Recently, new and improved models of inducible or constitutive mast cell deficiency were reported using Mcpt5-Cre mice crossed with simian diphtheria toxin transgenic mice (13). The Mcpt5-Cre system was also used to drive the expression of Kit^D814V in mature mast cells, leading to mastocytosis in the skin that increased with age and progressed to mast cell tumors and skin lesions (18). Crossing these mice with shp2^fl/fl mice would allow for testing the role of SHP2 in mastocytosis in vivo. We predict that SHP2 will be required for the survival of malignant mast cells and for the development of mastocytosis in mice. Future studies will be required to test whether SHP2 inhibitors can limit mast cell responses in allergy models or in systemic mastocytosis models.