Btk Plays a Crucial Role in the Amplification of FcϵRI-mediated Mast Cell Activation by Kit*

Stem cell factor (SCF) acts in synergy with antigen to enhance the calcium signal, degranulation, activation of transcription factors, and cytokine production in human mast cells. However, the underlying mechanisms for this synergy remain unclear. Here we show, utilizing bone marrow-derived mast cells (BMMCs) from Btk and Lyn knock-out mice, that activation of Btk via Lyn plays a key role in promoting synergy. As in human mast cells, SCF enhanced degranulation and cytokine production in BMMCs. In Btk-/- BMMCs, in which there was a partial reduction in the capacity to degranulate in response to antigen, SCF was unable to enhance the residual antigen-mediated degranulation. Furthermore, as with antigen, the ability of SCF to promote cytokine production was abrogated in the Btk-/- BMMCs. The impairment of responses in Btk-/- cells correlated with an inability of SCF to augment phospholipase Cγ1 activation and calcium mobilization, and to phosphorylate NFκB and NFAT for cytokine gene transcription in these cells. Similar studies with Lyn-/- and Btk-/-/Lyn-/- BMMCs indicated that Lyn was a regulator of Btk for these responses. These data demonstrate, for the first time, that Btk is a key regulator of a Kit-mediated amplification pathway that augments FcϵRI-mediated mast cell activation.

Mast cell activation leads to the release of both preformed and de novo synthesized inflammatory mediators. The intracellular signaling cascade regulating these responses is initiated by aggregation of high affinity receptors for IgE (Fc⑀RI) 4 following antigen binding to receptorbound IgE (1). However, antigen-induced triggering of mast cells in vivo is likely to occur with a background of stem cell factor (SCF)-mediated Kit activation, as SCF is essential for the growth, differentiation, homing, and survival of mast cells (2). By mimicking this situation in vitro, we have demonstrated that SCF dramatically augments both antigen-me-diated degranulation and cytokine generation in these cells (3,4). Kitmediated signals are thus required for optimal mast cell degranulation and cytokine production induced by Fc⑀RI aggregation.
Antigen-mediated degranulation and cytokine production are thought to be initiated by the activation of the Src family tyrosine kinase, Lyn (5). The resulting tyrosine phosphorylation of the ␤ and ␥ chains of Fc⑀RI promotes the binding of the tyrosine kinase Syk to Fc⑀RI (6). This permits the trans/auto-phosphorylation and activation of Syk (7,8), which in turn phosphorylates the transmembrane adaptor molecules LAT (9) and NTAL (3,10). These adaptor molecules orchestrate the recruitment of downstream signaling molecules to the receptor-signaling molecular complex by providing docking sites for cytosolic adaptor molecules, including SLP-76, Vav, Gads, Grb2, Gab1, and Gab2 (11) and signaling enzymes such as phospholipase (PL)C␥ 1 , PLC␥ 2 , and phosphoinositide (PI) 3-kinase (12,13). The subsequent elevation of intracellular calcium levels and activation of protein kinase C (PKC) leads to degranulation (14), whereas activation of the Ras-Raf-MAPK pathway induces arachidonic acid metabolite release (15) and downstream phosphorylation and activation of specific cytokine gene-related transcription factors (16). A parallel pathway controlled by the Src kinase, Fyn, also appears to help regulate Fc⑀RI-dependent mast cell activation (17).
Many of these same signaling events are initiated upon binding of SCF to Kit (18) but are insufficient on their own to induce degranulation (4). Our previous studies have suggested that this may be related to the inability of SCF to induce phosphorylation of LAT (3) and downstream activation of PKC (4). Nevertheless, SCF can potentiate Fc⑀RI-mediated degranulation and phosphorylation of NTAL as well as enhance calcium mobilization (3). How SCF augments these responses, however, was unclear. Given that the tyrosine kinase, Btk, is thought to play a role in the regulation of PLC␥-mediated calcium mobilization for both the B cell receptor (19) and the Fc⑀RI (20), we have examined whether Btk played a similar role in Kit-mediated responses. By use of bone marrowderived mast cells (BMMCs) from gene-deficient mice, Btk was not only found to be essential for the ability of SCF to potentiate antigen-mediated degranulation but was also found to be required for the ability of Kit to regulate cytokine production in antigen-stimulated cells.

EXPERIMENTAL PROCEDURES
Mast Cells-The Btk Ϫ/Ϫ , Lyn Ϫ/Ϫ , Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ knock-out, and wild type (WT) mice used in this study have been described previously (20). The mice were cross-bred on a C57BL/6 ϫ 129/Sv genetic background. The wild type mice were derived from the same parental lines as the knock-out mice. Breeding pairs heterozygous for Lyn, Btk (males were either BtkϪ/Y or ϩ/Y), or both were set up to generate both wild type and knock-out mice within the same litter. Whenever possible, littermates were compared directly. All animals were housed within the same room. The genotype of these mice was confirmed by reverse tran-* This work was supported in part by the NIAID Intramural Program of the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1  scription-PCR of tail biopsies and by immunoblot analysis of proteins extracted from the BMMCs derived from these mice. Bone marrow obtained by femur lavage was cultured in RPMI 1640 medium containing IL-3 as described (13). The studies were then conducted on BMMCs after 4 -6 weeks in culture.
Cell Activation-For degranulation and signaling studies, cultured BMMCs were sensitized overnight with anti-mouse monoclonal dinitrophenyl (DNP) IgE (100 ng/ml) (Sigma) in IL-3-free RPMI medium and then rinsed with HEPES buffer (21) containing 0.04% bovine serum albumin (Sigma). The cells were triggered in the same buffer with DNPhuman serum albumin (HSA) (0 -100 ng/ml) and/or murine SCF (0 -100 ng/ml; PeproTech, Rocky Hill, NJ) for 30 min for the degranulation studies or for the indicated periods for the signaling studies. For cytokine mRNA and release studies, cells were similarly sensitized but triggered for 4 or 10 h, respectively, in RPMI.
Degranulation Assay-Degranulation was monitored by the release of ␤-hexosaminidase into the supernatants (22). Briefly, BMMCs, sensitized as above, were triggered in 96-well plates (5 ϫ 10 5 cells per well, 100 l final volumes). The reactions were terminated by centrifugation (3000 rpm) at 4°C, and the supernatants were aliquoted to 96-well plates for ␤-hexosaminidase assay. The remaining cells were lysed by adding distilled water and freeze-thawing, and then aliquots were similarly assayed for ␤-hexosaminidase content. Degranulation was then calculated as the percentage of total (cells and supernatants) ␤-hexosaminidase content found in the supernatants following challenge.
Cytokine Production-RNase protection assays (RPA) were utilized to measure mRNA levels for multiple cytokines and chemokines following cell activation. Cells were sensitized and then triggered as above at a concentration of 10 ϫ 10 6 cells/ml. Messenger RNA was extracted by lysing the cells with 1 ml of TRIzol (Invitrogen) for 5 min at room temperature. Chloroform (200 l) was added to the lysates, and the mixtures were centrifuged for 15 min at 14,000 rpm. Isopropyl alcohol (500 l) was then added to the aqueous phases, and the mixture was incubated for 10 min to precipitate RNA. Ten g of RNA was used in the mRNA assay by using an in vitro transcription kit and pre-designed or custom-designed RPA templates (BD Biosciences). RPA was conducted according to the manufacturer's instructions; however, the synthesized radioactive probes labeled with [␣-33 P]UTP were purified with a probequant G-50 microcolumn (Amersham Biosciences) instead of ethanol precipitation, and the protected mRNA was precipitated with ethanol and ammonium acetate containing Glyco-blue (Ambion, Austin, TX). The gels were prepared with 80 ml of SequaGel-6 (National Diagnostics, Inc.), 20 ml of SequaGel-complete (National Diagnostics, Inc., Atlanta, GA), and 10% ammonium persulfate (Sigma). Levels of the secreted cytokines IL-4, IL-6, IL-13, and TNF-␣ were measured in the supernatants of activated BMMCs by ELISA (BIOSOURCE, Camarillo, CA).
Intracellular Calcium Determination-Calcium flux was measured in the BMMCs following loading of the cells with Fura-2 AM ester (Molecular Probes, Eugene, OR) as described (13). Cells were loaded with Fura-2 AM for 30 min at 37°C, rinsed, and resuspended in HEPES buffer containing 0.04% bovine serum albumin and sulfinapyrazone (0.3 mM) (Sigma), and then placed in a 96-well black culture plate (20,000 cells/well) (CulturPlat-96 F, PerkinElmer Life Sciences). Fluorescence was measured at two excitation wavelengths (340 and 380 nm) and an emission wavelength of 510 nm. The ratio of the fluorescence readings was calculated following subtraction of the fluorescence of the cells that had not been loaded with Fura 2-AM.

Kit Induces Lyn-dependent Phosphorylation of Btk in BMMCs
To investigate the role of Btk in Kit-mediated responses, we utilized BMMCs derived from the bone marrow of Btk Ϫ/Ϫ mice. Because previous studies had suggested that Btk and Lyn had both redundant and opposing functions in antigen-dependent mast cell (20) and B cell activation (24), we compared the responses in the Btk Ϫ/Ϫ BMMCs to those obtained in WT, Lyn Ϫ/Ϫ , and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ double knock-out BMMCs. The Btk Ϫ/Ϫ , Lyn Ϫ/Ϫ , and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMC genotypes were confirmed by probing lysates from these cells for Btk and Lyn (data not shown). The levels of expression of the other Tec kinases, including Tec and Itk (as controls for Btk) and other Src kinases, including Blk, Fgr, Fyn, Hck, c-Src, and Yes (as controls for Lyn), were unaffected in these cells, apart from a slight reduction in the expression of Fgr in the Lyn Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs (data not shown).
Both antigen and SCF induced the phosphorylation of Btk in WT mouse BMMCs (Fig. 1, a and b); however, maximum phosphorylation observed with SCF was of a lesser magnitude than that observed with antigen. Although there was little evidence of synergy in the responses at early time points (0 -120 s), when cells were co-stimulated with SCF and antigen, Btk phosphorylation was more sustained than was observed with the individual stimulants. As expected, this phosphorylation was not detected in the Btk Ϫ/Ϫ BMMCs (Fig. 1, c and d). In addition, the phosphorylation of Btk was substantially reduced in the Lyn Ϫ/Ϫ BMMCs indicating that the phosphorylation of Btk was largely dependent on Lyn.
Stimulation of BMMCs with antigen, but not SCF, resulted in an increase in the phosphorylation of the Src kinases (Fig. 1, e and f), although Src kinases were constitutively phosphorylated to some degree. In the Lyn Ϫ/Ϫ BMMCs there was virtually no phosphorylation of the Src kinases in both stimulated and non-stimulated BMMCs. Thus, the major Src kinase phosphorylated both constitutively and inducibly by antigen in the BMMCs was Lyn. However, overexposure of the gels revealed that SCF, but not antigen, also resulted in a lesser phosphorylation of another Src kinase that was not Lyn (Fig. 1e). There was little change in the phosphorylation of the Src kinases in the Btk Ϫ/Ϫ BMMCs, thus confirming that the phosphorylation of Btk is downstream of Lyn.

SCF Augments Fc⑀RI-mediated Degranulation and Cytokine Generation in Mouse BMMCs
To establish that SCF potentiated Fc⑀RI-dependent responses in WT mouse BMMCs as was the case in human mast cells (3,4), we examined degranulation and cytokine production in response to SCF, antigen, or both in combination. Fig. 2a shows that SCF, at concentrations up to 100 ng/ml, induced little degranulation. When added concurrently with antigen, however, SCF induced a marked concentration-dependent potentiation of antigen-mediated degranulation. Similarly, SCF and antigen acted in synergy to increase the message of multiple cytokines, including IL-1␣, IL-1␤, IL-4, IL-6, IL-13, TNF-␣, and interferon-␥ (Fig.  2b). To confirm that the potentiation of cytokine message levels translated into increases in cytokine protein, the release of TNF-␣, IL-6, and IL-13 was examined by ELISA 10 h following challenge with SCF with or without antigen. Again, as in human mast cells (4), cytokine secretion was minimally elevated in response to either SCF or antigen alone, but when added in combination, there was a marked synergistic enhancement of cytokine production (Fig. 2, c-e).  Cells were sensitized overnight with mouse monoclonal anti-DNP-IgE (100 ng/ml) in media without IL-3, and then following rinsing, the cells were challenged with SCF (100 ng/ml) or Ag (DNP-HSA, 100 ng/ml) or Ag and SCF added concurrently (100 ng/ml) for the indicated times (a and b), 120 s (c and d) or 30 s (e and f). Proteins were then extracted from the cells and, following separation by gel electrophoresis, were probed for the indicated phosphorylated proteins. e, the blots from the Lyn Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs were subsequently overexposed to reveal the Src kinases phosphorylated in response to SCF in the absence or presence of Ag. The blots are representative of n ϭ 3. b, the symbols used are as follows: •, Ag; OE, SCF; ࡗ, SCF and Ag concurrently. The data in b, d, and f were generated by scanning the blots in a, c, and e, respectively, and then normalizing to the maximal response obtained with antigen alone. d and f, the order of bars for each cell type is control (C), Ag, SCF, and AgϩSCF. The dashed line in f represents the constitutive phosphorylation in WT cells. DECEMBER 2, 2005 • VOLUME 280 • NUMBER 48  WT controls and was virtually abolished in the Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ double knock-out BMMCs (Fig. 3a). SCF was unable to potentiate the residual antigen-mediated degranulation (i.e. 10 -15%) in the Btk Ϫ/Ϫ and the Lyn Ϫ/Ϫ BMMCs and the minimal degranulation in the Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ double knock-out BMMCs (Fig. 3b). This is in contrast to SCF-mediated potentiation of degranulation induced by minimally effective concentrations of antigen as shown previously in Fig. 2a.

Btk in Kit-mediated Signaling in Mast Cells
Production of TNF-␣ and IL-6 and IL-13 (Fig. 3, c-e, respectively) in response to SCF and antigen or both in combination was reduced by ϳ50% in the Btk Ϫ/Ϫ BMMCs, potentiated in the Lyn Ϫ/Ϫ BMMCs, and virtually reduced to background levels in the Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ double knock-out BMMCs, as compared with WT BMMCs. In contrast to degranulation, however, an additive response to the combination of SCF and antigen was still observed in Btk Ϫ/Ϫ BMMCs, although the net response was still ϳ50% that in WT BMMCs. In the Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ double knock-out BMMCs, however, production of cytokines was virtually ablated. Similar responses were observed at the message level as determined by RPA (data not shown).

SCF-and Antigen-induced Signaling Studies in Btk ؊/؊ and Lyn ؊/؊ BMMCs
Activation of PLC␥ 1 and PI 3-Kinase-Our previous studies suggested that the ability of SCF to potentiate antigen-mediated degranulation was associated with an enhancement of calcium mobilization (4). As both PLC␥ 1 -and PI 3-kinase-dependent pathways control Fc⑀RImediated degranulation in human mast cells via regulation of calcium mobilization (13), we next examined whether these signaling events were ablated in the Btk Ϫ/Ϫ , Lyn Ϫ/Ϫ , and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs. PLC␥ 1 and PI 3-kinase activation was monitored by the phosphorylation of PLC␥ 1 or AKT, respectively (13). Both SCF and antigen stimulated PLC␥ 1 phosphorylation in WT BMMCs, and the combination of both stimuli resulted in an additive and more sustained phosphorylation than that induced by either stimulant alone (Fig. 4, a, b, e, and f). We were unable to detect PLC␥ 2 phosphorylation in response to SCF and antigen by utilizing a commercially available anti-phospho-PLC␥ 2 . However, following immunoprecipitation with an anti-PLC␥ 2 antibody and then probing with an anti-phosphotyrosine antibody, we observed that although both antigen and SCF induced PLC␥ 2 phosphorylation, these responses were not additive (data not shown). In contrast to PLC␥ 1 phosphorylation, the effects of SCF and antigen on AKT phosphorylation (Fig. 4, c, d, g, and h) were not additive. Rather, antigen induced a decrease in the more predominant SCF-mediated AKT phosphorylation (Fig. 4, c and d). This was likely because of Lyn-mediated down-regulation of PI 3-kinase activation, as the inhibitory response was reversed in the Lyn Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs but not the Btk Ϫ/Ϫ BMMCs (Fig. 4, g and h). The lack of synergistic enhancement of AKT phosphorylation in the Lyn Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs likely reflects the fact that both a Lyn-dependent inhibitory pathway, potentially via SHIP (25), and a Lyn-dependent activation pathway, potentially via Syk (26), for antigen-induced PI 3-kinase activation are blocked in these cells. This conclusion is further supported by the fact that the slight increase in AKT phosphorylation observed in response to antigen in the WT cells is absent in the Lyn Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs (Fig. 4, g and h) Btk-deficient BMMCs exhibited a slight reduction in PLC␥ 1 phosphorylation in response to antigen and SCF (Fig. 4, e and f). However, the ability of SCF to potentiate antigen-mediated PLC␥ 1 phosphorylation was completely blocked in the Btk Ϫ/Ϫ BMMCs, and the combination of antigen and SCF appeared to result in phosphorylation levels that were slightly lower than that observed with antigen alone (Fig. 4, e and f).
In Lyn Ϫ/Ϫ , as well as Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs, the ability of antigen in the absence or presence of SCF to induce PLC␥ 1 phosphorylation was completely blocked. As a result, the synergistic increase in PLC␥ 1 phosphorylation in response to SCF and antigen added concurrently was reduced to close to base line in the Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ double knock-out cells. Thus, although Lyn was required for phosphorylation of PLC␥ 1 in response to antigen, Btk was central to the ability of SCF to potentiate this response.  PLC␥ 1 (a, b, e, and f) and AKT (c, d, g, and h) in response to Ag, SCF, or SCF in the presence of Ag in WT (a-d), Btk ؊/؊ , Lyn ؊/؊ , and Btk ؊/؊ / Lyn ؊/؊ BMMCs (e-h). Cells were sensitized and treated with control buffer (C), Ag (DNP-HSA; 100 ng/ml), SCF (100 ng/ml), or Ag and SCF added concurrently (100 ng/ml) for the indicated times (a-d) or for 120 s (e-h), and proteins were extracted. Following gel electrophoresis, the proteins were probed with antibodies recognizing phosphorylated PLC␥ 1 (p-PLC␥ 1 ), phosphorylated AKT (p-AKT), or actin. The blots are representative of n ϭ 3-4. The data in b, d, f, and h were generated by scanning the blots in a, c, e, and g, respectively, and then normalizing to the maximal response obtained with antigen (b and f) or SCF (d and h) alone. The symbols used are as follows: •, Ag; OE, SCF; ࡗ, SCF and Ag concurrently in b and d. f and h, the order of bars for each cell type is control, Ag, SCF, and AgϩSCF. DECEMBER 2, 2005 • VOLUME 280 • NUMBER 48

JOURNAL OF BIOLOGICAL CHEMISTRY 40265
Calcium Mobilization-As with human mast cells (4), SCF and antigen acted in synergy to enhance calcium mobilization in WT BMMCs (Fig. 5a). In the Btk Ϫ/Ϫ BMMCs, the initial increases in calcium mobilization in response to antigen (Fig. 5b) or SCF (Fig. 5c) when added separately or concurrently (Fig. 5d) were still observed. These responses, however, were substantially lower and less sustained than those observed in the WT BMMCs. In contrast, in the Lyn Ϫ/Ϫ BMMCs, the increase in calcium levels was delayed but eventually reached levels that were similar to those in WT BMMCs. As was the case with degranulation, the residual calcium flux in the antigen-challenged Btk Ϫ/Ϫ (Fig.  5e) or Lyn Ϫ/Ϫ BMMCs (Fig. 5f) could not be further potentiated by SCF.
Taken together, the above data support the concept that the ability of SCF to potentiate antigen-mediated calcium mobilization, hence degranulation, was entirely dependent on Btk, and this was at the level of PLC␥ 1 activation but downstream of PI 3-kinase activation.
MAPK and Transcription Factor Phosphorylation-We next examined if the observed deficiencies in cytokine production in the Btk Ϫ/Ϫ , Lyn Ϫ/Ϫ , and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs correlated to reduced activation of MAPKs and specific transcription factors. In WT BMMCs, the phosphorylation of the ERK1/2, JNK, and p38 MAPKs was augmented by  /2 (a and b), p38 (c and d), and JNK (e and f) in response to Ag, SCF, or Ag and SCF added concurrently in WT, Btk ؊/؊ , Lyn ؊/؊ , and Btk ؊/؊ /Lyn ؊/؊ BMMCs. Cells were sensitized and treated with control buffer (C), Ag (DNP-HSA; 100 ng/ml), SCF (100 ng/ml), or Ag and SCF (100 ng/ml) for 30 min, and proteins were extracted. Following gel electrophoresis, the proteins were probed with antibodies recognizing phosphorylated ERK1/2 (p-ERK1/2), phosphorylated p38 (p-p38), or phosphorylated JNK (p-JNK), or actin. The blots are representative of n ϭ 3 experiments. The data in b, d, and f were generated by scanning the blots in a, c, and e, respectively, and then normalizing to the maximal response obtained with antigen alone. In these panels the order of bars for each cell type is control, Ag, SCF, and AgϩSCF.  5). e and f, the data from the WT BMMCs challenged with SCF and Ag from d, the Btk Ϫ/Ϫ or Lyn Ϫ/Ϫ BMMCs challenged with Ag and SCF concurrently from d, and the Btk Ϫ/Ϫ or Lyn Ϫ/Ϫ BMMCs challenged with Ag alone from b during the first 300 s have been re-plotted to demonstrate that the potentiation of antigen-mediated calcium flux by SCF is absent in the Btk Ϫ/Ϫ and Lyn Ϫ/Ϫ BMMCs. Standard errors in b, c, and d have been omitted for clarity but are similar to those in e and f. co-stimulation with antigen and SCF compared with the effects of the individual stimulants added alone (Fig. 6). There was no reduction in the synergistic phosphorylation of ERK1/2 in the Btk Ϫ/Ϫ and Lyn Ϫ/Ϫ BMMCs (Fig. 6, a and b) and only a slight reduction in the Btk Ϫ/Ϫ / Lyn Ϫ/Ϫ BMMCs. In contrast, the synergy between antigen and SCF in the phosphorylation of p38 MAPK and JNK was markedly impaired in all kinase-deficient BMMCs. Paradoxically, Lyn deficiency resulted in enhanced phosphorylation of both ERK1/2 and JNK in antigen-stimulated cells, and as a result, the additive effects of SCF on these responses were less apparent in these cells. Taken together, the above data indicate that the reduction in cytokine production in the Btk Ϫ/Ϫ and Btk Ϫ/Ϫ / Lyn Ϫ/Ϫ BMMCs was associated with similar deficiencies in the p38 and JNK signaling pathway(s) but not the ERK1/2 pathway. JNK and p38 regulate gene transcription by mediating the phosphorylation of transcription factors, including those of the AP1 complex (Fos and Jun), NFAT, and NFB (27)(28)(29). We thus examined these responses in the WT and kinase-deficient BMMCs. Both SCF and antigen induced the synthesis of Jun. However, these responses were not additive (Fig. 7, a and b). The subsequent phosphorylation of Jun in response to SCF and antigen added concurrently was greater than the responses of these agents when added separately (Fig. 7, c and d). We observed no marked differences in levels of Jun in the kinase-deficient FIGURE 7. Synthesis and/or phosphorylation of transcription factors in response to Ag, SCF, or Ag and SCF added concurrently in WT, Btk ؊/؊ , Lyn ؊/؊ , and Btk ؊/؊ /Lyn ؊/؊ BMMCs. Cells were sensitized and treated with control buffer (C), Ag (DNP-HSA; 100 ng/ml), SCF (100 ng/ml), or Ag and SCF (100 ng/ml) for 30 min, and proteins were extracted. Following gel electrophoresis, the proteins were probed with antibodies recognizing total (a, b, and e) and phosphorylated Jun (p-Jun) (c, d and e), total Fos (f and g), phosphorylated NFB (p-NFB) (h and i), and phosphorylated NFAT (p-NFAT) (j and k). The blots are representative of n ϭ 3 experiments. The data in b, d, g, i, and k were generated by scanning the blots in a, c, f, h, and j, respectively, and then normalizing to the maximal response obtained with antigen alone. The data in e were generated by normalizing the data in d with the data in b. In these panels, the order of bars is control, Ag, SCF, and AgϩSCF. BMMCs (Fig. 7, a and b) apart from slightly higher levels in the Lyn Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs challenged with SCF and antigen concurrently compared with the levels in WT BMMCs treated in a similar manner. In addition, no marked defects in the overall levels of phosphorylated Jun (Fig. 7, c and d) were observed in these cells compared with the WT responses. However, by taking into account the different levels of Jun expression in these cells (Fig. 7, a and b), there was a reduction in the phosphorylation of Jun per unit mass in response to SCF and antigen in combination in the Btk Ϫ/Ϫ , Lyn Ϫ/Ϫ , and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs (Fig. 7e), which reflected the observed changes in JNK phosphorylation described in Fig. 6, e and f. The lack of changes in the overall levels of phosphorylated Jun together with the relatively minor changes in Fos expression (Fig. 7, f and g) in these cells, however, suggest that the reduction in cytokine production in the Btk Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs was unlikely to be because of defects in the induction and/or phosphorylation of AP1 components.
Both NFB (Fig. 7, h and i) and NFAT (at the activating Ser-54) (Fig.  7, j and k) were phosphorylated in response to SCF and antigen in the WT BMMCs. The phosphorylation of NFB, in contrast to the phosphorylation of NFAT, was additive when these agents were added con-currently. The phosphorylation of NFB and NFAT, in response to SCF and antigen, added separately or concurrently, was partially reduced in the Lyn Ϫ/Ϫ BMMCs but substantially reduced in the Btk Ϫ/Ϫ and Btk Ϫ/ Ϫ/Lyn Ϫ/Ϫ BMMCs (Fig. 7, h-k). Taken together, these studies indicate that the defective phosphorylation of NFAT and NFB, rather than defective regulation of components of the AP1 complex, could be responsible for the reduced cytokine production in the Btk Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs.

DISCUSSION
Previous studies have focused on the role of Btk in regulating BCRmediated responses in B cells (24) and Fc⑀RI-mediated responses in mast cells (30 -32). However, the present studies point to a broader role of Btk as a core component of an amplification pathway that is utilized by Kit for augmenting the activation of mast cells via Fc⑀RI. As such, Btk may provide the link between receptor-proximal events and integrated downstream signaling by Kit and Fc⑀RI. These synergistic interactions likely mirror the situation that would be expected in vivo where SCF is essential for mast cell growth (33), differentiation (34), homing (35), and survival (36).
In the present studies, we have utilized Btk Ϫ/Ϫ and Lyn Ϫ/Ϫ mice to demonstrate the essential role for both kinases in the synergistic response to antigen and SCF in BMMCs. As shown previously in human mast cells (3,4), SCF was found to markedly potentiate degranulation and augment cytokine production in WT mouse BMMCs (Fig. 2). The enhancement of degranulation could be attributed to a synergistic increase in PLC␥ 1 phosphorylation and the resultant enhanced calcium mobilization. Examination of the kinetics of these and other signaling responses, including the phosphorylation of Btk (Fig. 1) and NTAL (data not shown), suggested that this may be because of an SCF-dependent conversion of the normally transient Fc⑀RI-mediated responses to a more sustained response. As reported previously (20), degranulation in response to antigen was reduced by ϳ50% in both Btk Ϫ/Ϫ and Lyn Ϫ/Ϫ BMMCs and was essentially absent in the Btk Ϫ/ Ϫ/Lyn Ϫ/Ϫ double knock-out BMMCs (Fig. 3). The additive defect in the Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs, compared with the single knock-out BMMCs, suggested that, although there is some overlap in the regulation and function of these enzymes, Btk and Lyn may also act independently to regulate degranulation (20).
The failure of SCF to enhance the residual antigen-induced degranulation in Btk Ϫ/Ϫ and Lyn Ϫ/Ϫ BMMCs indicated that both of these enzymes were essential for the ability of Kit to enhance mast cell degranulation. These observations correlated with deficient calcium signaling in the Btk Ϫ/Ϫ and Lyn Ϫ/Ϫ BMMCs (Fig. 5). Although the phosphorylation of Btk was partially regulated by Lyn, these enzymes appear also to have independent roles in mediating the calcium response. For example, the initial increase in the calcium signal observed in WT BMMCs in response to antigen was absent in the Lyn Ϫ/Ϫ BMMCs, whereas the initial increase in calcium flux in response to antigen and SCF was still evident in the Btk Ϫ/Ϫ BMMCs. However, this signal was less sustained resulting in substantially lower maximal calcium levels in these cells.
The inability of SCF to enhance the residual antigen-induced increase in the calcium signal in Btk Ϫ/Ϫ and Lyn Ϫ/Ϫ BMMCs was associated with an inability of SCF to augment the residual PLC␥ 1 phosphorylation in these cells (Fig. 4). Btk is known to activate PLC␥ by phosphorylating conserved activation tyrosine residues in the Src homology 2-Src homology 3 domain linker region in both PLC␥ 1 (Tyr(P)-771 and Tyr(P)-783) and PLC␥ 2 (Tyr(P)-753 and Tyr(P)-759) in B cells (37). We noted that although Lyn was absolutely required for the phosphorylation of Tyr(P)-783 in PLC␥ 1 in response to antigen in BMMCs, the phosphorylation of this residue was only partially reduced in the indi- We have restricted this scheme to signaling molecules actually examined in this study and excluded others, including cytosolic adaptor molecules and members of the Ras-Raf-MAPK-AP1 cascade for clarity. The readers are referred to Refs. 21, 27, and 28 for specific details about these molecules. In the diagram, the blue line represents the "principal" pathway for the activation of phospholipase C␥ 1 leading to degranulation and for the activation of the transcription factors NFB and NFAT leading to cytokine production. The green line represents the "amplification" pathway utilized by Kit for the potentiation of the principal pathway. The dashed lines with question marks represent unresolved stages of this cascade. In this scheme, Fc⑀RI aggregation results in the phosphorylation of the transmembrane adaptor molecules LAT and NTAL; however, Kit activation appears to result in the phosphorylation of NTAL in the absence of detectable LAT phosphorylation (4). The phosphorylation of LAT results in the recruitment and activation of PLC␥ 1 with resulting hydrolysis of membrane-bound phosphoinositide 4,5bisphosphate (PIP 2 ) to yield inositol trisphosphate (IP 3 ) and diacylglycerol (DAG). These molecules respectively induce mobilization of intracellular calcium and activation of PKC resulting in degranulation. The elevation of intracellular calcium levels also induces NFAT activation via the calcium-regulated phosphatase, calcineurin. Concurrently with these events, although slightly delayed in onset, is the activation of the amplification pathway that results in enhancement of PLC␥ 1 activity and calcium mobilization, and it is by this mechanism that we hypothesize that Kit potentiates Fc⑀RI-mediated degranulation and cytokine production. As previously described, this pathway appears to be regulated by NTAL (4) and, although not synergistically enhanced by SCF and antigen, by phosphoinositide 3-kinase (PI3K) (50). As both NTAL phosphorylation (4) and AKT phosphorylation (current study) are not abrogated in Btk Ϫ/Ϫ BMMCs, activation of Btk is likely downstream of these events, although the exact processes by which Btk is activated by this pathway is currently unclear. The activation of Btk by Kit, however, appears to be crucial for the subsequent potentiation of PLC␥ 1 -mediated calcium mobilization leading to degranulation and activation of NFAT and NFB leading to enhanced cytokine production.
vidual responses to antigen and SCF in the Btk Ϫ/Ϫ BMMCs. Nevertheless, Btk was absolutely required for the Kit-mediated enhancement of the phosphorylation of PLC␥ 1 (Tyr(P)-783) in response to antigen. These data suggest that in activated mast cells the phosphorylation of the critical Tyr(P)-783 in PLC␥ 1 is regulated both by a Lyn-dependent/ Btk-dependent pathway and a Lyn-dependent but Btk-independent pathway. It is the Lyn-dependent/Btk-dependent pathway, however, that is central to the amplification signaling cascade that is utilized by Kit to regulate Fc⑀RI-mediated mast cell degranulation. Our results, however, show some discrepancies with past observations in that it had been reported previously that antigen-dependent degranulation is either unchanged (38) or even enhanced (17,25) in Lyn Ϫ/Ϫ BMMCs. Nevertheless, our observations agree with reports of reduced antigendependent degranulation in these cells (20). The reasons for these apparent discrepancies remain unclear but may reflect different conditions for cell culture as, for example, the presence or absence of SCF.
The synergistic cytokine production in response to SCF and antigen (Fig. 2) in the BMMCs was accompanied by a marked synergistic phosphorylation of the MAPKs, ERK1/2, JNK, and p38 (Fig. 6) and a downstream synergistic phosphorylation of the transcription factors Jun and NFB (Fig. 7). Although the expression of Fos and Jun was increased by both antigen and SCF, additive or synergistic responses were not observed with the combination of stimuli. Likewise, synergy was not observed in the phosphorylation of the activating Ser-54 residue of NFAT. As reported for antigen-mediated production of IL-2 and TNF-␣ in BMMCs (20,39), elevated cytokine message and protein levels for multiple cytokines, including TNF-␣, IL-6, and IL-13, in response to SCF were reduced by about 50% in the Btk Ϫ/Ϫ BMMCs whether antigen was present or not (Fig. 3). The residual synergy observed in the Btk Ϫ/Ϫ BMMCs was essentially abolished in the Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ double knock-out BMMCs, at least for TNF-␣ and IL-13, despite an apparent enhancement in the Lyn Ϫ/Ϫ single knock-out BMMCs. The enhanced cytokine production in the Lyn Ϫ/Ϫ BMMCs may be because of a similar enhancement of ERK1/2 and JNK phosphorylation in response to antigen in these cells. Alternatively, this may reflect the reversal in the Lyn Ϫ/Ϫ BMMCs of the tonic inhibition of Kit-mediated PI 3-kinase by antigen that was observed in the WT BMMCs.
TNF-␣, IL-6, and IL-13 genes are regulated by binding of the transcription factors, NFAT and NFB (40 -42), and the AP1 complex to their promoter regions as regulated by AKT (32) and MAPKs, including JNK and p38 (43). The lack of defective AKT and ERK1/2 phosphorylation in Btk Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs indicates that the reduced cytokine production observed in these cells was not linked to these signaling molecules. However, the defects in cytokine production in the Btk Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs are accompanied by defective p38 and/or JNK signaling.
Of the downstream transcription factors examined, the only defects that correlated with the decreased cytokine production in the Btk Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs was the substantial reduction in the phosphorylation of NFAT and NFB. Despite significant reduction in JNK and p38 phosphorylation in these cells, we observed no apparent decreases in the synthesis and/or the overall total phosphorylation of AP1 components, Jun and Fos, although, on a per unit mass basis, a reduction in the phosphorylation of Jun in response to the combination of SCF and antigen was observed to reflect the changes in JNK phosphorylation. In T cells, NFAT is a target for p38 MAPK (43), and in B cells, NFAT activity is likely regulated by a Lyn-Syk-Btk-PLC␥ pathway through activation of the calcium-binding phosphatase, calcineurin (44 -46). It should be noted that although NFAT is regulated by dephosphorylation of multiple residues, its activity is also dependent on the calcium-and PKC-dependent phosphorylation of Ser-54 (47,48) examined in this study. Thus the deficiencies in NFAT phosphorylation in the Btk Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs may be explained by the reduced calcium signal and defective p38 phosphorylation in these cells. Similarly, both NFB and NFAT phosphorylations appear to be regulated by the converging activities of JNK and p38 in B cells (49). Again this might explain how defective phosphorylation of p38 and JNK would lead to similar defects in the phosphorylation of NFAT and NFB and ultimately reduced cytokine production in the Btk Ϫ/Ϫ and Btk Ϫ/Ϫ /Lyn Ϫ/Ϫ BMMCs.
As summarized in Fig. 8, in this paper we have presented data to support the conclusion that mast cell activation is regulated by both a primary and amplification signaling pathway and that Btk is an essential player in the amplification pathway that is utilized by Kit for the potentiation of mast cell mediator release. Moreover, these studies further reinforce the concept that allergic responses to antigen in a physiological setting must be viewed in the context of a background of Kit activation. Finally, these observations may set a novel paradigm for the way in which other stimuli such as adenosine, C3a, IL-3, IL-4, substance P, and chemokines, either induce or potentiate Fc⑀RI-mediated degranulation.