The Neurobeachin-like 2 Protein Regulates Mast Cell Homeostasis

The beige and Chediak–Higashi (BEACH) family members, such as neurobeachin-like 2 protein (Nbeal2), contain a BEACH domain, a concavalin A-like lectin (ConA) domain, WD40 domains, and pleckstrin homology domains (1). BEACH domains are crucial for vesicle transport (1) and fusion of vesicles with the cell membrane. ConA domains bind glycosylated proteins and are therefore important for protein sorting and secretion (1, 2), WD40 domains mediate protein–protein interaction (1) whereas pleckstrin homology domains interact with 4,5-bisphosphoinositol (3). In humans, loss-of-function mutations of NBEAL2 causes gray platelet syndrome (4–6), which is characterized by macrothrombocytopenia, a paucity of α-granules in megakaryocytes and platelets (7), and splenomegaly (8). Nbeal2 is highly expressed in cells of the hematopoietic system (4). However, alterations in immune reactions have not been studied in gray platelet syndrome patients. Mast cells contain granules and are typically activated by FcεRI cross-linking (9). FcεRI stimulation leads to NFAT activation, degranulation, the release of cytokines, chemokines, histamines, and proteases (10). Mast cells are effector cells of type I hypersensitivity, and are central to the pathogenesis of allergic diseases (11). Differentiation of mast cells depends on c-Kit and the IL-3R (12). Both receptors stimulate similar signaling pathways (e.g., STAT5 and PI3Ks), which are involved in mast cell proliferation and/or differentiation (13–15). Mast cells also express members of the TLR/IL-1 receptor family including IL-33R (16, 17), which mediates NF-κB–dependent (18) cytokine production but not degranulation (19). Given that mast cell effector functions depend on granules, and Nbeal2 is involved in granule biogenesis, we determined the functional relevance of Nbeal2 in mast cells.

We used sex- and age-matched (8–10 wk old) Nbeal2^−/− mice (20) and wild type (wt) littermates.

For genotyping PCR we used the forward (fw)-primer 5′-GTCCTGCTTGACCTACCGTC-3′, and the reverse (rw)-primers 5′-CAGGGAGGATAACGAGATAGTCTT-3′ (rw-primer 1), 5′-CCTAGGAATGCTCGTCAAGA-3′ (IRES-GT-Primer). Nbeal2^−/− mice were generated by targeting the exons 4 to 11 of the Nbeal2 gene. The exons 4–11 were replaced by an IRES element containing a selection cassette. The wt PCR with the fw-primer and the rw-primer 1 generates a 223 bp product (wt product). In contrast, the mutant PCR with the fw-primer and the IRES-GT-Primer generates a 401 bp product (mutant product).

For RT-PCR experiments, total RNA was isolated by using the peqGOLD TriFast kit (PEQLAB). The RNA was transcribed into cDNA by using the first-strand cDNA synthesis kit (Thermo Fisher Scientific) and random hexamer primers. Subsequently, cDNA was subjected to quantitative PCR by using the Maxima SYBR Green/ROX qPCR Master mix (Thermo Fisher Scientific). We used the fw-Nbeal2 primer 5′-TGTGAAGGGCTCTTTGACCC-3′ and the rw-Nbeal2 primer 5′-GGCCGGAGGGAACTTGTATT-3′ (both Sigma-Aldrich). For housekeeping genes we used the β2 microglobulin (fw-primer 5′-CTGACCGGCCTGTATGCTATC-3′ and rw-primer: 5′-TGCAGTCCCGCATAGTTGAA-3′, both Sigma-Aldrich). The quantitative PCR was performed by using ROCHE LightCycler 480.

Mice were treated intravenously with 3 μg rat anti-DNP IgE (D8406; Sigma-Aldrich) in a total volume of 100 μl PBS. After 20 h the body temperature was measured rectally. HSA-DNP (A6661; Sigma-Aldrich) (250 μg/100 μl PBS) was injected to induce anaphylaxis and body temperature was measured rectally. All animal experiments were approved by the appropriate institutional and governmental committees for animal welfare.

For generation of bone marrow–derived mast cells (BMMCs), bone marrow was obtained from the femurs and tibias of wt and Nbeal2^−/− mice. Bone marrow cells were cultured in IMDM (PAA) supplemented with 10% FCS, 100 U/ml penicillin, 100 mg/ml streptomycin, 50 mM 2-ME, and 20 ng/ml IL-3 (conditioned media from WEHI-3 cells). In the first week of BMMC generation, the medium was changed every second day and the adherent cells were discarded. In weeks 2 until 4 of culture, the medium was changed twice a week. After weeks 4 and 5, cell culture consisted of 95% BMMCs (identified by FcεRI and c-Kit). BMMCs were used for 4–6 wk after differentiation.

For analysis of the in vitro cultures, cells were blocked with anti-CD16/CD32 (clone 2.4G2) and rat-IgG (Jackson ImmunoResearch) and stained with PE-conjugated anti-murine c-Kit and FITC-conjugated anti-murine FcεRI (all BioLegend). For determination of mast cells in vivo, cells were blocked with anti-CD16/CD32 (clone 2.4G2) and rat-IgG (Jackson ImmunoResearch) and stained with FITC-conjugated anti-murine IL-33R (MD Bioproducts) and PE-conjugated anti-murine c-Kit (BioLegend). For stimulation experiments, BMMCs were left untreated, or were treated with PMA/ionomycin (Sigma-Aldrich), or primed with anti-DNP IgE (clone: SPE-7) (Sigma-Aldrich) overnight, and stimulated with HSA-DNP (Sigma-Aldrich). Subsequently, BMMCs were blocked with anti-CD16/CD32 (clone 2.4G2) and rat-IgG (Jackson ImmunoResearch) and stained with PacBlue-conjugated anti-murine CD107α, PE-conjugated anti-murine c-Kit, or FITC-conjugated anti-murine FcεRI (all BioLegend). For intracellular cytokine detection, BMMCs (4 × 10⁶ cells per ml) were left untreated or were treated with brefeldin A and stimulated with PMA/ionomycin. Subsequently, cells were fixed in 2% formaldehyde in PBS at room temperature for 20 min. For cytokine detection, fixed BMMCs were stained intracellularly with anti-CD117, and V450-conjugated anti-murine TNF-α (eBioscience) in 0.5% saponin, 0.5% BSA, and 2 mM EDTA. In all experiments cells were analyzed using an LSR II flow cytometer (BD) and FlowJo 9 (Tree Star, Ashland, OR).

BMMCs (10⁶ cells per ml) were IL-3 starved (1 h), stimulated with stem cell factor (SCF) or IL-3 (all from PeproTech), or were primed overnight with anti-DNP IgE (clone: SPE-7) and stimulated with HSA-DNP (both Sigma-Aldrich). In all experiments cells were lysed in buffer containing 20 mM HEPES (pH 7.5), 10 mM EGTA, 40 mM β-glycerophosphate, 2.5 mM MgCl₂, 2 mM orthovanadate, 1 mM DTT, 20 μg/ml aprotinin, 20 μg/ml leupeptin, and 1% Triton. For separation of the cytoplasmatic and nuclear fraction we used the ProteoJet-Kit (Fermentas) according to the experimental procedures. Protein concentration was determined by using the BCA-kit (Pierce). Protein samples were then boiled in 6 × Laemmli buffer.

Cell lysates were incubated with Abs against STAT3α, STAT5α, STAT5β, or c-Kit (all from Santa Cruz) (5 μg Ab/500 μg protein). We used mouse-, rabbit- or goat-IgG as nonspecific control Igs (Gentaur). Protein-G Sepharose (Invitrogen) was added and incubated for 4–6 h. Precipitates were washed and subsequently boiled in 6 × Laemmli buffer.

Samples were separated on SDS-Laemmli gels, using 8% (for detection of Nbeal2) or 10% poly acrylamide gels (for detection of phosphorylated/nonphosphorylated proteins), and were transferred onto nitrocellulose membranes (Biostep) by Western blotting. Membranes were blocked with dry milk and incubated with Abs detecting phosphorylated or nonphosphorylated proteins. We used anti–pY719-c-Kit, anti–pY694-STAT5, anti-STAT5, anti–pS463-p65, anti-p65, anti-pY525/pY526-Syk, anti-Syk, anti–pY783-PLCγ1, anti–pY191-Lat1, anti–pY171-Lat1, anti-Lat1, anti–pS176/pS180-IKK2, anti–pY705-STAT3, anti-STAT3, anti-MKK7, anti-MITF, anti-GATA2, anti-IRF8, anti-lamin, and anti-CD107α (all from Cell Signaling). Furthermore, we used anti–c-Kit, anti-STAT3α, anti-STAT5α, anti-STAT5β, anti-p27, anti–pY536-Shp1, anti-PLCγ1, anti-IKK2 (all Santa Cruz), anti-pNFAT, anti-tubulin (Sigma-Aldrich), and anti-Nbeal2 (Thermo Fisher Scientific) Abs. Membranes were washed in 0.1% Tween/TBS and incubated with HRP-conjugated secondary Abs: anti–rabbit-Ig, anti–goat-Ig (both Santa Cruz), and anti–mouse-Ig (Thermo Fisher Scientific). Detection was performed using ECL reagent (Pierce).

BMMCs (10⁶ cells per ml) were IL-3 starved, primed with SPE-7 (1 μg/ml) overnight, and stimulated with HSA-DNP (both Sigma-Aldrich). Furthermore, we stimulated with PMA and/or ionomycin (Sigma-Aldrich). Supernatants were analyzed by cytokine ELISAs with matched-paired Abs (BioLegend) or by the histamine ELISA (Histamine FAST ELISA; DRG, Springfield, NJ) according to the experimental procedures. For determination of the serum TNF-α and IL-6, we used ProcartaPlex Simplex Kit, High Sensitivity in combination with the ProcartaPlex Mouse High Sensitivity Basic Kit (Invitrogen) according to the experimental procedures. Serum histamine was determined by using the Histamine FAST ELISA (DRG) according to the experimental procedures.

BMMC were stimulated with IL-3 (20 ng/ml) (24 h), and were fixed in 70% ethanol (3 h; −20°C). Subsequently, cells were stained with a solution containing 2.5 mg/ml propidium iodide, 0.1 mg/ml RNase A, and 0.05% Triton X-100 (30 min). Cells were analyzed with the LSR II flow cytometer (BD) and FlowJo (Tree Star).

BMMCs (10⁶ cells per ml) were IL-3 starved (1 h), stimulated with IL-3 or SCF (both PeproTech), and cultured for further 54 h. Then [3H]-thymidine (1 μCi) in 25 μl complete IMDM (PAA) (without IL-3) was added to each well for an additional 18 h. Incorporated radioactivity was measured by using a β-scintillation counter (PerkinElmer). Shown is the stimulation index (SI), where the cpm values of wt BMMCs were set as one. The cpm values of stimulated BMMCs were the fold induction (SI) compared with unstimulated wt BMMCs.

BMMCs (5 × 10⁴ cells per μl) were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.2) and embedded in Epon. Sections were cut using an ultramicrotome (Leica Ultracut UCT), stained with 2% uranyl acetate (in 100% ethanol), lead in citrate, and examined with an EM900 transmission electron microscope (Carl Zeiss).

Ear pinna were fixed in 4% (v/v) buffered formalin and embedded in paraffin ensuring a cross-sectional orientation. For detection of mast cells, 5 μm sections were stained (Giemsa). Mast cells were counted per ear skin section over a length of 1 cm. Shown is the mean of all sections per genotype ± SD.

Western blots were quantified with ImageJ (National Institutes of Health). To quantify the expression of proteins, the intensities of protein bands were determined and normalized to the respective internal controls. To quantify protein phosphorylations, the intensities of the phosphorylation bands were determined and normalized to the respective total protein bands. The control (unstimulated cells) of wt cells was set as 1. We calculated the fold increase of phosphorylations compared with the unstimulated wt control.

Western blotting, flow cytometry, and PCR experiments were performed three times (one representative experiment is shown). Proliferation assays and ELISAs were performed three times (thereby, one experiment includes the BMMCs generated from n = 3 mice per genotype and two samples per mouse). Shown is the mean ± SD. For statistical analysis we used IBM SPSS Statistics (version 20.0; IBM, Ehningen, Germany). Statistical significance was assessed by unpaired Student t test. Statistical significance was accepted for p < 0.05.

To characterize the role of Nbeal2 in mast cells, we used in vitro–generated BMMCs from wt and Nbeal2^−/− mice. To determine if BMMCs express Nbeal2 we performed Nbeal2-specific PCRs. Additionally, we used a polyclonal Ab, which recognizes the C terminus of Nbeal2 (Supplemental Fig. 1A). Both the Ab and the PCRs detected Nbeal2 in wt but not in Nbeal2^−/− BMMCs (Supplemental Fig. 1B–D). Next, we investigated the influence of Nbeal2 in in vitro BMMC differentiation in the presence of IL-3. Compared to wt bone marrow cultures, in Nbeal2^−/− bone marrow cultures the fraction of BMMCs (c-Kit⁺/FcεRI⁺) was increased in the first but decreased in the second week. After 3 wk, the fraction of BMMCs was similar in bone marrow cultures from wt and Nbeal2^−/− mice (Fig. 1A, 1B). In contrast, the total cell numbers and absolute BMMC numbers (Fig. 1C, 1D) were decreased in Nbeal2^−/− cultures throughout the whole culture period. We also tested the in vitro differentiation in the presence of IL-3 in combination with SCF, where we found similar results as shown in the presence of IL-3 alone (Supplemental Fig. 1E–H). These data indicate that Nbeal2 is not critical for mast cell differentiation but mediates the expansion of BMMC in vitro. Mast cell expansion is regulated by the MAPK-kinase MKK7 and cell-cycle inhibitors such as p27 (21). Therefore, we tested the expression of both proteins in lysates of wt and Nbeal2^−/− bone marrow cultures. As shown in Fig. 1E, Nbeal2 deficiency resulted in increased levels of MKK7 and p27 (for statistical analysis of the Western blots, see Supplemental Fig. 1I). This indicates that the reduced BMMC numbers in the Nbeal2^−/− bone marrow cultures result from the increased expression of MKK7 and p27.

GATA2 and MITF regulate the expression of p27 (22–24). Furthermore, the expression of GATA2 and MITF is mediated by IRF8 (13, 25, 26). Therefore we hypothesized that the increased protein level of p27 results from increased levels of IRF8, GATA2, and MITF. Indeed, the protein levels of IRF8, GATA2, and MITF are increased in Nbeal2^−/− compared with wt BMMCs (Fig. 1E; for statistical analysis of the Western blots, see Supplemental Fig. 1I). In Nbeal2^−/− cultures the increased MITF protein level persisted for 5 wk, whereas the protein levels of IRF8 and GATA2 were upregulated in the third but downregulated in the fourth and fifth week (Fig. 1E; for statistical analysis of the Western blots, see Supplemental Fig. 1I). These data indicate that NBeal2 negatively regulates the IRF8-MITF/GATA2-p27 pathway.

Next, we compared the mast cell numbers in wt and Nbeal2^−/− mice in vivo. To do this, we investigated the peritoneal fluid and the ear skin. Compared to wt mice, the mast cell numbers in the peritoneal fluid (Fig. 1F, 1G) and in ear skin (Fig. 1H, 1I) were reduced in Nbeal2^−/− mice. Taken together, these data indicate that Nbeal2 controls the number of mast cells in vitro and in vivo.

We found that Nbeal2 deficiency results in overexpression of transcription factors and of the cell-cycle inhibitor p27 in BMMC cultures. Therefore, we speculated that Nbeal2 is a nuclear protein that controls the cell cycle and proliferation. However, the Nbeal2 sequence does not contain nuclear localization sequences, and we only detected Nbeal2 in the cytoplasm but not in the nucleus of BMMCs (Supplemental Fig. 2A, 2B), excluding nuclear localization. Next, we investigated the cell-cycle progression, and the SCF- and IL-3–induced proliferation of mature wt and Nbeal2^−/− BMMCs. Compared to wt BMMCs, in Nbeal2^−/− BMMCs we found a cell-cycle arrest (Fig. 2A), and a reduced proliferation rate (Fig. 2B, 2C). Thus these data indicate that Nbeal2 controls cell-cycle progression and BMMC expansion in vitro.

Next, we investigated the effects of Nbeal2 deficiency on the SCF- and IL-3–induced STAT signaling, which is critically involved in mast cell differentiation and proliferation (13, 14, 27). Nbeal2 deficiency did not influence the SCF-induced c-Kit activation but resulted in a reduced SCF- (Fig. 2D, 2E; for statistical analysis of the Western blots, see Supplemental Fig. 3A–C) and IL-3– (data not shown) induced STAT3 and STAT5 activation. STATs are negatively regulated by Shp1 in mast cells (27). Therefore, we investigated the SCF-induced activation of Shp1 in wt and Nbeal2^−/− BMMCs. Compared to wt BMMCs, in Nbeal2^−/− BMMCs we found a strongly increased SCF-induced Shp1 activation (Fig. 2F; for statistical analysis of the Western blots see Supplemental Fig. 3D), indicating that Nbeal2 negatively regulates Shp1 and thus suppresses the activation of STAT5.

During these Western blot analyses we also found that Nbeal2 is expressed as the 300 kDa and additionally in a 200 kDa variant in BMMCs (Fig. 2F). Thereby, the 200 kDa variant is destabilized in response to SCF (Fig. 2F). Together, our data show that Nbeal2 regulates the activation of Shp1, STAT3, and STAT5, and is expressed as a 300 kDa and a yet-uncharacterized 200 kDa variant in mast cells.

Because Nbeal2 mediates effective STAT activation, we hypothesized that Nbeal2 interacts with STATs. In STAT3 (Fig. 3A) and STAT5 (Fig. 3B) precipitates we predominantly detected the 200 kDa Nbeal2 variant indicating complex formation between these proteins. Given that Nbeal2 interacts with STATs and regulates their SCF-induced activation, we hypothesized that Nbeal2 deficiency might influence the interaction between c-Kit and STATs. Unexpectedly, we could not detect Nbeal2 in c-Kit precipitates but confirmed that c-Kit interacts with STAT3/5 in wt BMMCs (Fig. 4). In Nbeal2^−/− BMMCs, the interactions between c-Kit and STAT3/5 were strongly increased (Fig. 4). Together, these data show that Nbeal2 does not directly interact with c-Kit but controls the assembly of the c-Kit/STAT signalosome.

Next, we investigated FcεRI-mediated signaling, the major receptor system pivotal to mediate mast cell effector functions. FcεRI signaling leads to a Syk-dependent Lat1 phosphorylation (28) and therefore to membrane localization and to activation of the PLCγ1 (28–30). Furthermore, Lat1 phosphorylation is critical for MAP-kinase activation (31). Compared to wt BMMCs the activation of Syk was strongly, and of the PLCγ1 slightly increased in Nbeal2^−/− BMMCs (Fig. 5A; for statistical analysis of the Western blots see Supplemental Fig. 3E, 3F). However, the phosphorylation of the Syk substrate Lat1 (Fig. 5A; for statistical analysis of the Western blots see Supplemental Fig. 3G, 3H), and the activation of IKK2, p65, JNK1/2, p38, and of Erk1/2 (Fig. 5B; for statistical analysis of the Western blots see Supplemental Fig. 4A–E) was similar in wt and Nbeal2^−/− BMMCs. This demonstrates that Nbeal2 controls the FcεRI-mediated Syk and PLCγ1 activation in BMMCs.

In platelets, Nbeal2 controls the release of cytokines (32) by regulating the biogenesis of α-granules (5, 20). Mast cell functions also depend on granules. Therefore, we investigated whether Nbeal2 mediates granule formation and thus mast cell effector functions. Flow cytometry and transmission electron microscopy showed that Nbeal2^−/− BMMCs are smaller and less granular than wt BMMCs (Fig. 5C–E), but express the same amount of the FcεRI and of the IL-33R (Supplemental Fig. 4F). Next, we determined whether Nbeal2 regulates the fast mast cell response induced by the FcεRI by investigating degranulation, and the release of preformed cytokines. To trigger the FcεRI, BMMCs were primed with the DNP-specific IgE Ab, SPE-7, and were stimulated with HSA-DNP for 1 h. Afterwards, we investigated the surface expression of CD107α, which correlates with degranulation (33), and examined the release of preformed cytokines such as MCP-1 and TNF-α (34, 35), as well as the production of IL-6 and IL-13. In Nbeal2^−/− BMMCs, the CD107α surface expression (Fig. 5F) and the release of MCP-1, TNF-α, and IL-6 were increased (Fig. 5G–I). In contrast, FcεRI stimulation did not induce an IL-13 release from wt or Nbeal2^−/− BMMCs (Supplemental Fig. 4G), indicating that IL-13 is not stored in granules as a preformed cytokine.

The increased surface expression of CD107α and the increased release of cytokines indicated enhanced degranulation, despite fewer granules in Nbeal2^−/− BMMCs. To test this hypothesis, we compared the total amount of CD107α and the histamine release in wt and Nbeal2^−/− BMMCs. Compared to wt BMMCs, in Nbeal2^−/− BMMCs, the total amount of CD107α was strongly increased (Fig. 5J). In contrast, the histamine release was decreased in Nbeal2^−/− BMMCs (Fig. 5K).

Furthermore, we investigated the late mast cell response, which depends on the de novo cytokine synthesis. Therefore, we stimulated the FcεRI for 4–8 h. Nbeal2 deficiency resulted in an increased release of MCP-1, TNF-α, and IL-6, whereas the release of IL-13 was similar in wt an Nbeal2^−/− BMMCs (Fig. 6A–D).

Next, we wanted to determine whether Nbeal2 regulates the synthesis of cytokines. Thus we investigated the intracellular content of cytokines in response to stimulation with PMA/ionomycin. Similar to the FcεRI stimulation, PMA/ionomycin treatment also induced an increased CD107α surface expression, and increased the release of TNF-α in Nbeal2^−/− BMMCs (Fig. 6E, 6F). Furthermore, the intracellular content of cytokines was also increased in Nbeal2^−/− BMMCs (Fig. 6G). Together, these data indicate that Nbeal2 mediates granule formation, and negatively controls the cytokine production.

To investigate the role of Nbeal2 in IgE-mediated anaphylaxis, we sensitized mice with SPE-7 and treated them with HSA-DNP. We did not detect differences in the IgE-mediated anaphylaxis between wt and Nbeal2^−/− mice (Fig. 6H). This result was confirmed by the similar content of histamine, IL-6, and TNF-α in the serum of wt and Nbeal2^−/− mice in response to FcεRI-mediated anaphylaxis (Fig. 6I). These results support the hypothesis that the decreased mast cell number caused by Nbeal2 deficiency is compensated by the hyper-reactive phenotype, which is characterized by an increased production of cytokines in Nbeal2^−/− mast cells (Fig. 7).

Nbeal2 mediates granule formation in megakaryocytes and platelets (5, 20, 36). We found that Nbeal2 is also critical for biogenesis of granules and therefore for storage and release of cytokines in mast cells. Deppermann et al. (20) showed that Nbeal2 is not involved in megakaryocyte differentiation, but controls the proliferation of megakaryocytes. We investigated the role of Nbeal2 in mast cell differentiation and proliferation. We found that Nbeal2 deficiency leads to reduced mast cell numbers in vivo. It is possible that the number of mast cells might be influenced by the surrounding environment, resulting in fewer mast cells in vivo. However, in in vitro cultures the exogenous influence of surrounding cells is minimized, and in Nbeal2^−/− cultures the number of BMMC is also strongly reduced, instead indicating intrinsic defects of Nbeal2^−/− mast cells. Intrinsic defects could be caused by an altered expression of transcription factors or of cell-cycle inhibitors in Nbeal2^−/− BMMCs. Mast-cell differentiation is driven by the IRF8-dependent expression of GATA2 (25, 26) and the PI3K-MITF signaling axis (13). Furthermore, the expansion of mature mast cells is regulated by cell-cycle inhibitors such as p27 (21). Interestingly, in Nbeal2^−/− bone marrow cultures, the protein levels of IRF8, MITF, GATA2, and p27 are strongly increased. Given that the expression of p27 depends on GATA2 (23) and MITF (24), we conclude that the increased protein level of the IRF8-GATA2/MITF transcription factor cascade results in an increased expression of p27 in mature mast cells. Consequently, the increased expression of p27 induces cell-cycle arrest and therefore a blockade of the growth factor–induced proliferation of mature mast cells. Together, our data indicate that Nbeal2 is not important for mast cell differentiation but controls their proliferation. Therefore, Nbeal2 maintains appropriate numbers of mature mast cells in the periphery by regulating the amount of IRF8, GATA2, MITF, and finally of p27. How Nbeal2 regulates the amount of these proteins is unknown. Nbeal2 is not located in the nucleus, excluding a direct influence of Nbeal2 on the transcriptional machinery. We hypothesize an essential role of Shp1 and STAT5 in the Nbeal2-mediated regulation of IRF8, GATA2, MITF, and p27 as well as in the resulting mast cell proliferation. The SCF-induced STAT5 activation and proliferation is reduced in Nbeal2^−/− BMMCs. We suppose that strongly activated Shp1 is causative for the reduced STAT5 activity, and the decreased proliferation rate of Nbeal2^−/− BMMCs. This is underpinned by the fact that 1) overexpression of wt Shp1 leads to inhibition of STAT5 (27); and that 2) either overexpression of wt Shp1 or a dominant negative STAT5 mutant results in strongly reduced BMMC proliferation (27).

Thus, we hypothesize that the reduced STAT5 activation results in increased expression of IRF8 in Nbeal2^−/− BMMCs. Interestingly, Esashi et al. (37) demonstrated that STAT5 negatively regulates IRF8 expression. Consequently, the Shp1-mediated STAT5 suppression in Nbeal2^−/− BMMCs leads to upregulation of IRF8, MITF/GATA2, and p27, and finally to cell-cycle arrest and the blockade of proliferation. This indicates that dysregulated Nbeal2 expression or gain-of-function mutations of Nbeal2 result in altered activation of STATs and thus in altered expression patterns of several transcription factors.

This model fits the concept that Nbeal2 variants interact with STATs, and mediate their effective activation. Unexpectedly, STATs predominantly interact with a novel and yet uncharacterized short 200 kDa Nbeal2 variant in BMMCs. We screened the Uniprot database but could not find a murine Nbeal2 variant with a molecular mass of 200 kDa. All listed short Nbeal2 splice variants (e.g., the Nbeal2 variant with the UniProt Accession number F6VTL9; http://www.uniprot.org/uniprot/F6VTL9) lack the ConA domain which contains the N-terminal part. Therefore, we predict a 200 kDa Nbeal2 variant, which also lacks the N-terminal part but is identical to the C-terminal part of the 300 kDa Nbeal2 variant (UniProt Accession number E9Q9L6; http://www.uniprot.org/uniprot/E9Q9L6) (Fig. 7A). However, the structural organization of the 200 kDa Nbeal2 variant and the specific roles of the 300 kDa and the 200 kDa variant are unknown, but are under investigation in our laboratory. Nevertheless, we conclude that Nbeal2 interacts with STATs and controls the assembly of the c-Kit/STAT signalosome and its regulation by Shp1 (Fig. 7B, 7C).

Importantly, these NBeal2 functions are not limited to the growth factor–induced signaling. The FcεRI-triggered signaling is also partly regulated by Nbeal2. Thereby, Nbeal2 negatively regulates the activation of Syk and PLCγ1. The FcεRI-triggered Lat1 phosphorylation and the resulting downstream signaling (e.g., the activation of JNK1/2, Erk1/2, and p38) is described as a Syk-dependent event (28, 31). However, we found increased Syk activation, but neither increased phosphorylation of Lat1 nor increased activation of JNK1/2, Erk1/2, or p38. This indicates uncoupled and Syk-independent signaling in Nbeal2^−/− BMMCs. Therefore, we hypothesize that Nbeal2 controls the localization and therefore the activation of Syk in mast cells.

In addition to the regulation of signaling pathways, Nbeal2 also controls compartmentalization and effector functions of mast cells. In Nbeal2^−/− BMMCs, the numbers of granules and therefore the release of histamine were reduced. In contrast, the expression of CD107α is increased. This indicates that Nbeal2^−/− BMMCs have fewer granules, which contain more CD107α compared with wt BMMCs. Furthermore, in Nbeal2^−/− BMMCs we also found an increased production and release of cytokines, as well as an increased amount of GATA2. GATA2 is essential for cytokine production in mast cells (38). Thus we conclude that the increased cytokine production is mediated by the increased amount of GATA2. Therefore, the reduced number of granules and thus the decreased storage capacity is compensated for by enhanced production of preformed or de novo–synthesized cytokines. How Nbeal2 regulates granule formation and the release of preformed or de novo–synthesized mediators in mast cells remains unknown. Together, these results demonstrate that Nbeal2 mediates granule biogenesis and controls the production of cytokines in mast cells.

Unexpectedly, the fast FcεRI-mediated anaphylactic reaction, which strongly depends on histamine-producing mast cells (11, 39), was not altered in Nbeal2^−/− mice. This result is confirmed by similar histamine and cytokine levels in wt and Nbeal2^−/− mice in response to the anaphylactic reaction. Together, these data indicated that the lower mast cell number is compensated for by an increased responsiveness of the remaining mast cells in Nbeal2^−/− mice. Indeed, in vitro–generated Nbeal2^−/− BMMCs produce more cytokines than wt BMMCs, indicating an increased responsiveness of Nbeal2^−/− BMMCs. However, in contrast to the release of cytokines, in vitro–generated Nbeal2^−/− BMMCs release less histamine compared with wt BMMCs. The discrepancy in the histamine release might be explained by mast cell extrinsic factors, which might compensate for the reduced histamine response of mast cells in NBeal2^−/− mice. Thereby, endothelial cells can be excluded because they do not express Nbeal2, and show normal synthesis and storage of several factors in Nbeal2^−/− mice (20, 40). We rather speculate that Nbeal2 influences the differentiation of other cell lineages (e.g., basophils), which are also important for anaphylactic reactions and therefore might compensate the reduced histamine release from mast cells in Nbeal2^−/− mice.

Together, we conclude that the reduced mast cell number might be compensated by 1) an increased responsiveness of the remaining mast cells; and/or by 2) an altered differentiation of other cell lineages in Nbeal2^−/− mice. In summary, we show that Nbeal2 is a multifunctional protein regulating mitogenic signaling, cell-cycle arrest, and mast cell effector functions. Together with a recent report that shows that a homozygous nonsense NBEAL2 mutation can mimic the clinical symptoms of the autoimmune lymphoproliferative syndrome (41), our data support the relevance of Nbeal2 in immunity above and beyond granule biosynthesis.

We thank Dr. F.-D. Böhmer (Centre of Medical Biomedicine, Jena) for critical reading.

The authors have no financial conflicts of interest.