I787 provides signals for c-Kit receptor internalization and functionality that control mast cell survival and development

Mast cells (MCs) are potent effector cells that play a key role in allergic disorders, autoimmunity, inflammation, and protective immune responses against bacteria and parasites.^1-5  MCs differentiate from multipotent hematopoietic progenitors in the bone marrow and migrate to connective and mucosal tissues where maturation mainly occurs.^6,7  MC numbers are increased at the site of inflammation and regulated via proliferative and antiapoptotic mechanisms. Signaling through the membrane receptor tyrosine kinase c-Kit is essential for both human and mouse MC development, survival, maturation, and activation.⁷  Furthermore, stem cell factor (SCF), the specific c-Kit ligand, acts as the major chemotactic factor for MCs.⁸ 

The c-Kit receptor is a 975-amino acid protein composed of an extracellular ligand-binding domain which contains 5 immunoglobulin (Ig)-like domains and a single transmembrane domain. The cytoplasmic region of c-kit contains a split protein tyrosine kinase (PTK) domain with an insert of 77 hydrophilic amino acids. The crystal structure of c-Kit revealed that SCF interacts with the first, second, and third Ig-like domains.^9,10  Upon ligand binding, the fourth and fifth Ig domains of 2 neighboring c-Kit ectodomains are brought closer, stabilizing the interaction between 2 receptor molecules.¹⁰  The cytoplasmic region carrying the protein tyrosine kinase domain displays a bilobed architecture with a small N-terminal lobe (amino acids 582-671), and a large C-terminal lobe (amino acids 678-953).^11-13  The cleft generated between the 2 lobes contains the kinase catalytic site. Bivalent binding of SCF to c-Kit leads to receptor dimerization, autophosphorylation, creates binding sites for SH2-containing phosphotyrosine-binding proteins, and induces kinase activation. A large number of signaling molecules has been identified as either interacting or as substrate proteins for c-Kit. These include the p85 subunit of phosphatidylinositol 3′ kinase (PI3K), phospholipase Cγ (PLCγ), Src family tyrosine kinases, and the adaptor proteins Grb2, Grb7, and APS.^14-16  PI3K-mediated activation of protein kinase B (PKB, Akt), the Ras-mitogen–activated protein kinase (MAPK) cascade, the Jak-Stat pathway, and Ca²⁺ signaling are examples of c-Kit downstream signaling cascades affecting multiple cellular functions.

c-Kit mutations have been identified in the membrane proximal Ig-like domain (exon 8 and 9), in the juxtamembrane domain (exon 11), and in the tyrosine kinase domain. Naturally occurring mouse mutants for c-Kit and mouse strains generated by site-directed mutagenesis indicate that most of the mutations that affect MC development are loss-of-function mutations. In humans, loss-of-function mutations in c-Kit result in piebaldism,¹⁷  an autosomal dominant disease characterized by a white forelock and large, nonpigmented patches on the forehead, eyebrows, chin, chest, abdomen, and extremities corresponding with a loss of MCs in affected skin areas.¹⁸  However, several point mutations in the tyrosine kinase, juxtamembrane, or transmembrane domain lead to ligand-independent constitutive activation of c-Kit and systemic mastocytosis.^19,20 

Here we describe a novel point mutation, an A to T transversion at position 2388 of the c-Kit gene (at the Pretty2 allele also known as m2Btlr), which results in the substitution of isoleucine 787 (I787) by phenylalanine, and analyze the consequences of this amino acid substitution for c-Kit signaling and MC development.

C57BL/6J mice were used for N-ethyl-N-nitrosurea (ENU) mutagenesis to generate the Kit^787F/787F (Kit^787F) strain. Four- to 6-week-old mice were analyzed. The Kit^787F mutation was observed in G1 mice born to ENU mutagenized sires (pure C57BL/6J background) and mapped as a dominant phenotype by outcrossing an affected animal to C3H/HeN and outcrossing a second time to the same strain. On 34 meioses, the mutation was confined to a critical region bounded by D5Mit352 and D5Mit158. This interval contained the Kit gene, which was sequenced. The stock was maintained by crossing for 6 generations to C57BL/6J. Genetically MC-deficient WBB6F₁-Kit^W/Kit^W-v (Kit^W/Kit^W-v) and control WBB6F₁-Kit^+/+ (Kit^+/+) mice were bred in the Animal Care Facility of the Research Center Borstel and housed in specific pathogen–free facility. All experiments were performed according to institutional guidelines. Eight-week-old Kit^W/Kit^W-v mice were reconstituted with 3 × 10⁶ wild-type or Kit^787F bone marrow–derived mast cells (BMMCs) injected intraperitoneally. Six weeks after reconstitution peritoneal cells (PECs) were analyzed for the presence of MCs by fluorescence-activated cell scanning (FACS).

PECs were isolated by peritoneal lavage with 10 mL of cold 0.9% NaCl solution. BMMCs were generated by cultivation of bone marrow cells from Kit^787F, Kit^787F/+, and control C57BL/6 mice in the presence of recombinant murine interleukin-3 (IL-3). Cells were maintained in complete medium consisting of 10% heat-inactivated fetal calf serum (FCS; Biochrom), 50μM β-mercaptoethanol (Invitrogen), nonessential amino acids, 2mM l-glutamine, penicillin, streptomycin, and 1mM sodium pyruvat (all from Invitrogen) in Iscove modified Dulbecco medium (IMDM; PAA) supplemented with 5 ng/mL IL-3 (R&D Systems). After 4 weeks of culture, BMMCs represented > 98% of the total cells according to FACS analysis of cell-surface expression of T1/ST2 and FcεRI and were negative for CD11c, B220, F4/80, Gr-1, and CD3 expression. BA/F3 cells (pro-B-cell line) were cultivated in RPMI 1640 (Invitrogen) supplemented with 10% FCS and 10 ng/mL IL-3.

To analyze cytokine production, BMMCs (2 × 10⁶/mL) were stimulated with SCF (R&D Systems), IgE (SPE-1; Sigma), dinitrophenylated human serum albumin (DNP-HSA; Sigma), lipopolysaccharide (LPS; Salmonella enterica serovar Friedenau, kindly provided by H. Brade, Research Center Borstel, and purified as previously described²¹ ) or phorbol-12-myristate-13-acetate (PMA) and Ionomycin (both from Sigma-Aldrich) for 48 hours at 37°C and 5% CO₂. Supernatants were collected, and cytokines and chemokines were measured by enzyme-linked immunosorbent assays (ELISAs) using specific antibodies and standard proteins from R&D Systems. To analyze BMMC proliferation in response to IL-3, 2 × 10⁶ cells/mL were incubated with IL-3 for 72 hours, and 1 μCi ³H-thymidine was added for the last 18 hours. Measurement of histamine by o-phthaldialdehyde assay,²²  measurement of MC degranulation by β-hexosaminidase release,²³  RNA extraction, cDNA synthesis, and conventional semiquantitative reverse transcription polymerase chain reaction (RT-PCR) for MC proteases are described in the supplemental data (available on the Blood Web site; see the Supplemental Materials link at the top of the online article).

Sequencing of cDNA and genomic DNA, and in silico protein analysis are described in the supplemental data. To clone mouse c-Kit (CD117, c-Kit, accession number of NM_001122733) into the entry vector pDONR221, the coding sequence was first amplified from a mouse cDNA library using the forward primer 5′-CACCATGAGAGGCGCTCGCGGCGCCTGGG-3′ and the reverse primer including a c-Kit-stop codon 5′-TCAGGCATCTTCGT GCACGAGCAGGGG-3′. After sequencing of the 2944-bp insert, a c-Kit-pDON221 clone was recombined with Invitrogen pDEST-40 expression vector according to the manufacturer's instructions. c-Kit cDNA-mutant (2388A→T) was generated by PCR amplification of 2 c-Kit fragments overlapping at the mutation site. Briefly, c-Kit upstream fragment of 2395 bp was amplified from cDNA library using the forward primer 5′-CACCATGAGAGGCGCTCGCGGCGCCTGGG-3′ and the reverse primer carrying the mutation 5′-GGCTGCCAAATCTCTGTGAAAACAATTCTTGGAGGCGAGGA-3′. The c-Kit downstream fragment of 590 bp was amplified from cDNA library using the forward primer carrying the mutation 5′-TCCTCGCCTCCAAGAATTGTTTTCACAGAGATTTGGC AGCC-3′ and the reverse primer with the c-Kit stop codon 5′-TCAGGCATCTTCGTGCACGAGCAGGGG-3′. Both fragments were purified and mixed in a final PCR to generate the c-Kit coding sequence of 2944 bp carrying the A→T mutation resulting in an exchange of amino acid I787 to phenylalanine. The cloning procedure of the c-Kit mutant into the expression vector pDEST40 was as described above for the c-Kit wild-type cDNA.

To transiently transfect BA/F3 cells, the Cell Line Nucleofector kit V (Lonza) was used. For transfection, 2 × 10⁶ cells were washed and resuspended in 100 μL of Nucleofector solution V together with the 6 μg DNA. Cells were pulsed in the Nucleofector I Device (Lonza) and subsequently transferred into a 6-well plate containing 2 mL of culturing medium. Twenty-four hours after the transfection, the transient c-Kit expression was measured by flow cytometry.

To obtain stable clones, the cells were selected for 2 weeks with 600 μg/mL G-418 (PAA). For this purpose, 24 hours after transfection, cells were diluted out into 96-well plates from a starting density of 3 × 10³ cells/mL.

Cells were washed twice in FACS buffer (2% newborn calf serum, 0.1% NaN₃, 2mM ethylenediaminetetraacetic acid [EDTA] in phosphate-buffered saline [PBS]) and staining with phycoerythrin (PE)–, allophycocyanin (APC)–, or fluorescein isothyocyanate (FITC)–conjugated antibodies against cell-surface molecules. Monoclonal antibodies against CD117 (c-Kit, 2B8), CD16/32 (Fcγ III/II, 2.4G2), CD54 (3E2), CD81 (Eat2; all from BD PharMingen), T1/ST2 (DJ8; Morwell Diagnostics), CD53 (OX-80), Serotec, FcεRI (MAR-1; eBioscience) were used for surface staining. To prevent nonspecific binding, all samples were preincubated with Fc-Block or unlabeled isotype matched nonspecific antibodies (BD PharMingen). Samples were analyzed on a FACSCalibur flow cytometer (BD Biosciences) according to standard protocols. Gates on viable cells were set according to exclusion of propidium iodide (PI) staining. To assess SCF binding, cells were washed with PBS/5% FCS and preincubated on ice for 2 hours. Recombinant mouse SCF (R&D Systems) was added, cells were washed twice and then incubated with biotinylated goat α-SCF (R&D Systems) following the incubation with streptavidin (SA)-PE (Dianova). All incubation steps were performed on ice, all washing steps at 250g at 4°C. To measure ligand-receptor internalization, cells were preincubated for 2 hours on ice, then 50 ng/mL SCF were added, cells were incubated on ice for 1 hour, and then washed twice. Internalization was induced by incubation of cells at 37°C, stopped by the addition of ice-cold PBS/5%FCS, and SCF was measured on the cell surface as described above. Inhibition of internalization was performed by preincubation of cells with 100μM Dynasore [3-hydroxynaphthalene-2-carboxylic acid (3,4-dihidroxybenzylidene) hydrazide; Tocris Bioscience] at 37°C for 30 minutes. Receptor internalization was induced and measured as described above. Dynasore was added in all incubation steps.

Apoptosis of BMMCs and stably transfected BA/F3 cells were measured by annexin V FITC kit (Bender MedSystems) according to the manufacturer's instructions.

Tissue samples were fixed with 4% formaldehyde, Carnoy fixative or Bouin fixative for 24 hours, and then processed for paraffin embedding. Cytospins were prepared by centrifugation of the PECs on glass slides by cytocentrifuge (Shendon). Tissue sections or cytospins were stained with toluidine blue or naphtolesterase (Sigma-Aldrich). Staining for MC protease-2 (MCP-2) was performed by MCP-2–specific antibodies²⁴  and donkey rabbit-specific secondary antibodies conjugated with horseradish peroxidase (HRP) and 3,3′-diaminobenzidine (DAB) as a substrate.

Cell pellets were lysed for 15 minutes on ice in cell extraction buffer: 1% NP-40, 50mM Tris-HCl buffer, pH 8.0, 150mM NaCl, 10mM sodium fluoride, 1mM sodium orthovanadate, and protease inhibitors (complete protease inhibitor cocktail [Roche] or 1 μg/mL pepstatin A, 1 μg/mL leupeptin, 10mM phenylmethylsulphonylfluoride [PMSF]). The detergent-insoluble materials were removed by centrifugation for 15 minutes at 13 000 rpm (16 000g) at 4°C. For immunoprecipitation studies, cell lysates were incubated with 0.5-2 μg primary antibody at 4°C overnight followed by addition of 20 μL of Protein A/G beads (Calbiochem) and further incubation for 1 to 2 hours. Immune complexes were collected and washed 4 times with cell extraction buffer. Proteins from cell lysates and from immunoprecipitates were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and detected by Western blotting using specific antibodies. Equal cell equivalents were subjected to SDS-PAGE. For some experiments, protein concentrations were determined using a bovine serum albumin (BSA) protein assay kit (Bio-Rad Laboratories) and 50 μg proteins were analyzed by electrophoresis in SDS-PAGE. Antibodies against phospho-Erk1/2, phospho-p38, p38, phospho-PLCγ1, PLCγ1, phospho-c-Kit (Tyr719), c-Kit, PKB (Akt), phospho-p90Rsk, phospho-Stat5 (Tyr694), and phospho-IKKβ were purchased from Cell Signaling Technology. Antibodies against Erk1/2, phospho-Cbl (Tyr700), ubiquitin (P4D1), and actin were purchased from Santa Cruz Biotechnology. Tyrosine-phosphorylated proteins were detected with p-Tyr-100 mouse monoclonal antibodies (Cell Signaling Technology).

IgE-preloaded BMMCs (0.15 μg/mL IgE, overnight) were washed with RPMI, resuspended at 5 × 10⁶ cells/mL in RPMI containing 1% FCS, 30μM Indo-1 AM (Invitrogen), and 0.045% pluronic F-127 (Invitrogen) and incubated for 45 minutes at 37°C. Cells were then pelleted, resuspended in RPMI containing 1% FCS, and analyzed in an LSR II (Becton Dickinson) after the indicated stimulation procedures. The FACS profiles were converted to line graph data using the FlowJo application Version 8.8.6 (TreeStar).

Statistical analysis of the results was performed by Student t test for unpaired samples. A P value lower than .05 was considered statistically significant.

Kit^787F mice generated by ENU mutagenesis of C57BL/6J mice were identified by their distinct coat color pattern. Heterozygotes (Kit^787F/+) have a diluted coat color, light ears, and tail, and white spots on the belly and back. Homozygous Kit^787F mice have black eyes and a white coat color, indicating a lack of functional melanocytes, which strongly resembles mice carrying a mutation at the white spotting (W) locus, the c-Kit locus on the mouse chromosome 5 (supplemental Figure 1A). In addition, Kit^787F mutant mice are infertile, anemic, and show substantial hematopoietic abnormalities (Z.O., M.S., S.B.-P., manuscript in preparation). Therefore, the c-Kit gene was sequenced, and a mutation was identified. Five fragments covering the annotated full-length transcript (accession number NM021099) were amplified from cDNA and directly sequenced by dye terminator sequencing. Genomic DNA was subsequently used to sequence the 21 exons of the gene, including the splice sites. Only 1 mutation was found in mice carrying 1 or 2 mutant alleles corresponding to an A to T transversion at position 2388 of the Kit transcript (Figure 1A). The 787F mutation results in the substitution of I787 by phenylalanine in the C-lobe of the kinase domain (Figure 1A and supplemental Figure 1B), which leads to a change from an aliphatic to an aromatic amino acid in the vicinity of the proton acceptor aspartate at position 790. The protein residue I787 is highly conserved across many vertebrate species (supplemental Figure 1C) from puffer fish to humans, suggesting an essential role of I787 for c-Kit activation and/or signaling.

Since mutations in c-Kit are known to affect MC development, we tested whether the 787F mutation has an effect on MC differentiation in vivo. As shown in Figure 1, MCs were absent in Kit^787F and markedly reduced in Kit^787F/+ in peritoneum (2.8% ± 0.7% of peritoneal cells; n = 7 in wild-type vs 1.5% ± 0.4%; n = 6 in Kit^787F/+), as demonstrated by toluidine blue staining and FACS analysis (Figure 1B,E), ear tissue, peripheral lymph nodes, and skin stained by toluidine blue (Figure 1C and supplemental Figure 2A-B), trachea and stomach stained for chloroacetate esterase activity (Figure 1D and supplemental Figure 2C), and spleen, trachea, and lymph nodes stained for mouse MCP-2 (mMCP-2; supplemental Figure 2D and data not shown). These data demonstrate that the 787F mutation leads to the loss-of-function of c-Kit and a complete developmental block of both mucosal and connective tissue-type MCs in vivo.

To investigate the effect of the 787F mutation on MC differentiation in vitro, bone marrow cells of wild-type and Kit^787F mice were cultured in the presence of IL-3. BMMCs were generated, and expression of characteristic surface receptors, histamine content, protease expression, and proliferation as well as signaling in response to IL-3 were analyzed. Wild-type and Kit^787F BMMCs display a comparable expression of the high affinity IgE receptor, FcεRI, the IL-33 receptor ST2, CD16/32 (Fcγ III/II receptor), tetraspanins CD53 and CD81, and adhesion molecule CD54/ICAM-1 (supplemental Figure 3A). Furthermore, histamine content and expression of MC proteases (supplemental Figure 3B-C), as well as IL-3–induced proliferation and signaling (supplemental Figure 4A-B; Stat5, Erk1/2, and PKB phosphorylation) were also comparable between Kit^787F and wild-type BMMCs. These data suggest that IL-3–driven BMMCs differentiation in vitro is not affected by the c-Kit 787F mutation.

To study whether the 787F mutation modulates c-Kit expression, wild-type and Kit^787F BMMCs were analyzed by FACS using c-Kit–specific monoclonal antibodies. The mean fluorescence intensity (MFI) after staining with c-Kit–specific antibodies was slightly reduced in Kit^787F BMMCs compared with wild-type BMMCs (244 ± 41; n = 7 in wild-type BMMCs vs 158 ± 10; n = 9 in Kit^787F BMMCs; P = .038; Figure 2A). However, Kit^787F BMMCs do not display any significant difference in the ligand binding ability, measured as MFI of BMMCs preincubated with SCF and stained with α-SCF antibodies (Figure 2B). Because SCF is a dimer and binds with high avidity to the receptor dimer, one can conclude that both mutant and wild-type c-Kit receptors dimerize at a comparable level of efficiency.

A mechanism that is well known to control receptor signaling is shedding. Because tumor necrosis factor α-converting enzyme (TACE)-mediated cleavage of c-Kit generates a soluble receptor form that blocks SCF activities in vitro and in vivo,^25-27  we investigated whether c-Kit shedding in 787F BMMCs is affected. Interestingly, c-Kit receptor shedding induced by PMA was comparable between wild-type and Kit^787F BMMCs (Figure 2C). Thus, despite the slightly reduced c-Kit expression in Kit^787F BMMCs, the 787F mutation does not affect ligand binding and receptor shedding.

To investigate the functional consequences of the 787F c-Kit mutation, the ability of SCF to prevent MC apoptosis induced by growth factor deprivation was tested. As shown in Figure 3A, IL-3 withdrawal induced a strong and comparable apoptosis in Kit^787F and wild-type BMMCs. However, the addition of SCF in the culture medium rescued wild-type, but not Kit^787F BMMCs from cell death in a dose-dependent manner.

Next, the ability of in vitro differentiated Kit^787F BMMCs to reconstitute MC-deficient mice and to survive in vivo was tested. Kit^787F BMMCs and wild-type BMMCs used as controls were injected intraperitoneally into MC-deficient WBB6F₁-Kit^W/Kit^W-v (Kit^W/Kit^W-v) mice, and peritoneal MC numbers were determined by FACS 4 weeks later (Figure 3B). Contrary to wild-type BMMCs, Kit^787F BMMCs did not reconstitute the peritoneal MC population in MC-deficient mice, indicating that c-Kit–mediated survival signals are essential and nonredundant for long-lasting MC survival in vivo.

Next, the production of proinflammatory cytokines, such as IL-6, was tested in Kit^787F and Kit^787F/+ BMMCs and compared with wild-type BMMCs upon SCF stimulation. SCF neither induced IL-6 release in Kit^787F BMMCs (Figure 3C) nor enhanced the FcεRI-mediated cytokine production (Figure 3D), whereas Kit^787F/+ BMMCs showed a 75% decrease in IL-6 production (supplemental Figure 5A). In contrast, cytokine production induced by LPS or PMA and ionomycin was comparable between Kit^787F and wild-type BMMCs (Figure 3E-F). Furthermore, degranulation of BMMCs in response to IgE and antigen (AG) was comparable between wild-type, Kit^787F/+ and Kit^787F BMMCs (supplemental Figure 5B), but an SCF-mediated degranulation enhancement could be detected only in wild-type BMMCs (data not shown). Thus, the 787F mutation in BMMCs specifically affects SCF-mediated cytokine production and SCF-driven rescue from apoptosis.

To assess the effects of the 787F mutation on c-Kit–mediated signal transduction, we stimulated BMMCs with SCF and monitored the activation of various signaling parameters. SCF stimulation rapidly induces the phosphorylation of Erk1/2 and p38 in BMMCs. While at early times (1 minute) after SCF treatment the activation of these MAPKs was comparable in wild-type, Kit^787F/+, and Kit^787F BMMCs, SCF-induced Erk1/2 and p38 activation was decreased in Kit^787F/+ BMMCs and impaired at later time points in Kit^787F BMMCs (Figure 4A and supplemental Figure 5C). In wild-type cells, phosphorylation of Erk1/2 and p38 was maintained for at least 15 to 30 minutes, whereas MAPK activation was hardly detectable in Kit^787F BMMCs after 5 minutes of SCF stimulation (Figure 4A). Interestingly, phosphorylation of p90Rsk, a downstream target of Erk1/2, was strongly induced in wild-type MCs, but largely abolished in Kit^787F MCs (Figure 4B), indicating that the short and transient activation of Erk1/2 in Kit^787F BMMCs was not sufficient for efficient downstream signal propagation.

A similar failure of Kit^787F BMMCs to sustain SCF-mediated signaling was observed for the SCF-induced activation of PKB (supplemental Figure 6B). Signal transduction was not generally altered in Kit^787F BMMCs, as FcεRI-mediated activation of Erk1/2, p38, PKB, IKKβ, and PLCγ1 as well as calcium influx was normal in Kit^787F BMMCs (Figure 4C, supplemental Figure 6B,D, and data not shown).

c-Kit-induced activation of PKB is mediated via PI3K and is thought to depend on the recruitment of the regulatory p85 subunit of PI3K to phosphorylated Y719 of c-Kit. Analysis of tyrosine phosphorylation revealed marked impairment of c-Kit Y719 phosphorylation in SCF-stimulated Kit^787F BMMCs compared with wild-type BMMCs (Figure 4B). Only at 30 seconds after SCF stimulation was autophosphorylation of 787F mutant c-Kit detectable, although at strongly reduced levels compared with wild-type cells (supplemental Figure 6A). Consistent with significantly attenuated autophosphorylation of 787F c-Kit, only small amounts of p85 were found in c-Kit immunoprecipitates of SCF-stimulated Kit^787F BMMCs (supplemental Figure 6C). Calcium mobilization, mediated through PLCγ activation and regulated by PI3K, is quickly induced by SCF.²⁸  PLCγ1 phosphorylation in response to stimulation with SCF was hardly detectable in Kit^787F BMMCs, but was rapidly induced in wild-type BMMCs (Figure 4B). Concomitantly to the strongly reduced activation of PLCγ1, SCF-induced calcium signaling was also significantly attenuated in Kit^787F BMMCs, although an initial influx of calcium could still be observed (Figure 4D). Recently it has been demonstrated that c-Kit stimulation also induces the phosphorylation and activation of the ubiquitin E3 ligase Cbl, which mediates c-Kit ubiquitination and receptor internalization.^29,30  While in wild-type cells, Cbl phosphorylation was observed as early as 1 minute upon SCF-stimulation, Cbl activation was hardly detectable in Kit^787F BMMCs (Figure 4B). Moreover, immunoprecipitation experiments clearly demonstrated c-Kit ubiquitination in response to SCF-stimulation in wild-type cells, which was largely defective in Kit^787F BMMCs (Figure 4E). Thus, 787F c-Kit does not efficiently mediate Cbl activation and subsequent receptor ubiquitination, suggesting that c-Kit internalization may be defective in Kit^787F BMMCs.

As additional proof of our findings, wild-type and mutant c-Kit cDNA were cloned and stably transfected in BA/F3 cells (BA/F3 c-Kit and BA/F3 c-Kit^787F, respectively). BA/F3 c-Kit or BA/F3 c-Kit^787F cells were stimulated with SCF at different time points, and SCF-mediated Erk1/2 phosphorylation was investigated. Despite expressing high levels of c-Kit on the cell surface, as measured by anti-c-Kit FACS staining (Figure 5A), the BA/F3 c-Kit^787F cells did not show any Erk1/2 phosphorylation (Figure 5B). Furthermore, as shown in Figure 5C, only BA/F3 c-Kit and not BA/F3 c-Kit^787F are successfully rescued from cell death in response to SCF, while IL-3 is equally effective in mediating cell survival. This proves that the substitution of I787 by phenylalanine in c-Kit is responsible for the defect in SCF-mediated c-Kit signaling.

Next, to test whether the 787F mutation affects the internalization of the ligand-receptor complex (SCF-c-Kit), Kit^787F and wild-type BMMCs were preincubated with SCF and subsequently stained with anti-SCF–specific antibodies. As shown in Figure 6A, while after 30 minutes of incubation, 63% of bound SCF was internalized by wild-type BMMCs, Kit^787F BMMCs internalized only 1% of the bound SCF. A 60% internalization of the SCF-c-Kit complex by Kit^787F BMMCs required more than 2 hours of incubation. Even after 4 hours, SCF was still detectable on the surface of Kit^787F BMMCs. Comparable results were obtained by detection of c-Kit surface expression levels by flow cytometry over time (supplemental Figure 7). These results indicate that despite comparable SCF binding to c-Kit and normal receptor shedding, the internalization of the SCF-c-Kit complex is impaired in BMMCs carrying the 787F mutation.

To investigate whether inhibition of c-Kit endocytosis in SCF-stimulated wild-type BMMCs affects c-Kit downstream signaling, wild-type BMMCs were treated with Dynasore, a cell-permeable dynamin inhibitor (Figure 6B-C). The c-Kit internalization was tested by FACS analysis (Figure 6B) and was found to correlate, with slightly different dynamics, with the reduction of Erk1/2 phosphorylation and c-Kit autophosphorylation as determined by Western blot analysis (Figure 6C). Therefore, in wild-type BMMCs, c-Kit endocytosis is required for c-Kit kinase activity supporting the concept that c-Kit signaling is regulated via receptor internalization.

Previous studies have provided evidence that loss-of-function mutations in the murine dominant white spotting/c-kit locus affect MC differentiation and function. We have now identified a new c-Kit loss-of-function mutation that confers severe MC deficiency: Kit^787F mice lack both mucosal and connective tissue-type MCs. The 787F mutation in BMMCs affects SCF-mediated functions such as cytokine production and apoptosis rescue. Furthermore, the known synergistic effect of SCF on FcεRI-mediated MC activation is impaired by this c-Kit mutation. Interestingly, c-Kit downstream signaling in Kit^787F BMMCs is initiated normally, but fails to be sustained thereafter. Finally, we provide evidence that the mechanism responsible for the observed defect in MC differentiation and survival is the inhibition of c-Kit receptor internalization, phosphorylation, and ubiquitination. Together, these observations conclusively demonstrate that I787 is essential in regulating c-Kit function and, thereby, MC development and survival. Our study, thus, not only identifies yet another crucial site in the c-Kit receptor protein, whose mutation has grave consequences for MC biology, but also introduces a novel concept namely that sustained c-Kit signaling and receptor internalization are required for MC development, activation, and survival.

A large number of mutations in white spotting locus (W) allelic to c-Kit have been identified. Between them, mutations located upstream the kit transcription start to affect the expression of Kit to a different degree as well as MCs numbers in vivo.^31,32 W^sh mice bear an inversion of ∼3.1 Mb inactivating transcriptional regulatory elements 67.5 kb upstream of the c-Kit transcription start,^32,33  leading to tissue-specific dysregulation of c-Kit expression. In W^sh mice, all organs lack MCs, with the exception of skin, in which MC numbers decrease in an age-dependent manner.³⁴  The WBB6F1-Kit^W/Kit^W-v mouse strain, which is frequently used to study MC functions in vivo and regarded as MC-deficient, is heterozygous for 2 distinct c-Kit mutations—a T660M substitution (Kit^W-v allele) and a deletion mutation of 78 amino acids resulting in the absence of the transmembrane domain (Kit^W allele), leading to a block in MC development. F856S substitution within the tyrosine kinase domain in Wads mutant strain is reported to lack MCs specifically and only in the skin.³⁵  Furthermore, amino acid substitutions impairing receptor autophosphorylation (eg, Y567F and Y569F at the juxtamembrane domain) have demonstrated the essential function of these residues in activation of downstream signaling pathways in vitro and MC development in vivo.³⁶  However, only combined mutations of both tyrosine residues resulted in a severe deficiency in SCF-mediated signaling and loss of MCs in the peritoneal cavity, stomach, and dermis, while the single mutations Y719F, Y569F, and Y567F lead to a strong reduction of MC numbers in the peritoneum, but not skin.^36-38  In contrast to the above reported mutations, our study reports that the single 787F point mutation alone is sufficient for the complete absence of both connective tissue types and mucosal MCs and underlines the essential regulatory role of this residue in MC development. 787F mutation does not affect c-Kit shedding, induced by phorbol esters,³⁹  indicating that wild-type and 787F c-Kit are comparable in the accessibility for sheddases, which are in charge of cleaving c-Kit. Because induction of CD53, CD54/ICAM expression and histamine production are induced in BAF3 cells transfected with c-Kit bearing the activating mutation D816V,⁴⁰  we asked whether these markers are down-regulated in 787F BMMCs. However, our data suggest that expression of CD53, CD54/ICAM-1, and histamine do not correlate with inhibitory effects of 787F mutation. Further, we could observe that the expression of MC proteases is not affected by 787F mutation.

The 787F mutation in BMMCs not only affects SCF-mediated functions such as cytokine production and cell survival, but also impairs the synergistic effect of the SCF-c-Kit signaling pathway to FcεRI-mediated signals.

I787 resides in the C-lobe of the kinase and immediately precedes the catalytic loop of the kinase (consisting of the sequence HRDLAARN), which surrounds the actual site of phosphate transfer from ATP to the substrate and could be important for appropriate phosphorylation and tyrosine kinase activity. In addition, the crystal structure of activated c-Kit indicates that I787 could form hydrogen bonds with R815 of the C-lobe, which may serve to stabilize the activation loop in an extended conformation required for kinase activation.¹²  Interestingly, the 787F mutation does not seem to disrupt the catalytic loop, since early signaling events such as Erk1/2 and p38 phosphorylation take place normally after ligand binding in Kit^787F BMMCs. However, sustained signaling is impaired, and receptor internalization is severely decelerated. Therefore our data raise the possibility that receptor internalization is an essential step for sustained c-Kit signaling events, rather than inactivation of ligand-receptor complexes, and that receptor internalization and autophosphorylation are so intimately intertwined that both kinase activity and receptor endocytosis may well operate side-by-side.

Furthermore, the MAPK pathway and the PI3K pathway are reported as distinct pathways regulating proliferation and survival. Namely, they are activated independently of each other downstream of c-Kit in MCs.^41,42  However, our study demonstrates that the 787F mutation affects both pathways temporally in a similar manner, both being activated immediately after ligand-mediated engagement of the receptor and impaired in Kit^787F BMMCs at later time points.

That the substitution of I787 by phenylalanine in c-Kit is responsible for the defect in SCF-mediated c-Kit signaling was also proven in BA/F3 cells transfected with mutant c-Kit cDNA. BA/F3 c-Kit^787F cells show an inhibited Erk1/2 phosphorylation and were not rescued from cell death in response to SCF. Because c-Kit internalization in transfected BA/F3 cells, in comparison to primary MCs takes place at low rate (data not shown), these cells were not suitable to investigate c-Kit internalization dynamics.

Recent studies on signal compartmentalization indicate that receptor endocytosis plays a greater role in cellular biology than merely serving as a degradation pathway. Receptor internalization targets activated receptors to endocytic compartments, where specific signaling complexes can be assembled and/or stabilized.^43,44  While MAPK activation by receptor tyrosine kinases is thought to be initiated at the plasma membrane, it has been demonstrated that endosomes support sustained MAPK activation.^45,46  Therefore, in future studies it will be essential to examine in detail which of the c-Kit-induced signaling pathways are dependent on receptor internalization.

Finally, identification of small molecule mimetics of the I787 region could be a novel approach for developing c-Kit antagonists that are of importance for the treatment of patients with aberrant c-Kit tyrosine kinase activity, but do not target the conventional kinase active site. This appears all the more important since the D816V mutant of c-Kit frequently observed in systemic mastocytosis patients is resistant to treatment with STI-571 (Gleevec, imatinib). Targeting c-Kit may also be relevant in MC-associated disorders in which SCF acts as an amplifier of antigen-mediated cell activation.

We thank Manuel Hein, Kathrin Westphal, Katrin Streeck, Gesine Rohde, and Marlies Kaufmann for excellent technical support, Elena Bulanova for helpful discussion, and Nora Smart for critical reading of the manuscript.

This work was supported by the Deutsche Forschungsgemeinschaft (DFG; SFB/TR22 A14 to S.B.P and HU794/4-2 to M.H). P.R. is supported by Clusters of Excellence “Future Ocean” and “Inflammation at Interfaces.”

Contribution: Z.O., N.F., M.H., J.M., F.M., M.S., P.R., and A.B. performed the research; X.D. and B.A.B. contributed the mice; T.G. contributed to the histologic analyses; and Z.O., N.F., and S.B.P. analyzed data and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Silvia Bulfone-Paus, Department of Immunology and Cell Biology, Research Center Borstel, Parkallee 22, D-23845 Borstel, Germany; e-mail: sbulfone@fz-borstel.de; or Zane Orinska, Department of Immunology and Cell Biology, Research Center Borstel, Parkallee 22, D-23845 Borstel, Germany; e-mail: zorinska@fz-borstel.de.