Degranulation and Anaphylaxis Efficient Ige-Mediated Mast Cell

The Tetraspanin CD63 Is Required for Efficient IgE-Mediated Mast Cell Degranulation and Anaphylaxis

This information is current as Stefan Kraft, Marie-Hélène Jouvin, Nitin Kulkarni, Sandra of September 26, 2021. Kissing, Ellen S. Morgan, Ann M. Dvorak, Bernd Schröder, Paul Saftig and Jean-Pierre Kinet J Immunol 2013; 191:2871-2878; Prepublished online 14 August 2013;

doi: 10.4049/jimmunol.1202323 Downloaded from http://www.jimmunol.org/content/191/6/2871

Supplementary http://www.jimmunol.org/content/suppl/2013/08/14/jimmunol.120232 Material 3.DC1 http://www.jimmunol.org/ References This article cites 48 articles, 20 of which you can access for free at: http://www.jimmunol.org/content/191/6/2871.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 26, 2021

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2013 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

The Tetraspanin CD63 Is Required for Efﬁcient IgE-Mediated Mast Cell Degranulation and Anaphylaxis

Stefan Kraft,*,† Marie-He´le`ne Jouvin,* Nitin Kulkarni,* Sandra Kissing,‡ Ellen S. Morgan,x Ann M. Dvorak,x Bernd Schro¨der,‡ Paul Saftig,‡ and Jean-Pierre Kinet*

Mast cell (MC) activation through the high-affinity IgE receptor Fc«RI leads to the release of mediators involved in immediate-type allergic reactions. Although Abs against the tetraspanins CD63 and CD81 inhibit Fc«RI-induced MC degranulation, the intrinsic role of these molecules in Fc«RI-induced MC activation is unknown. In MCs, CD63 is expressed at the cell surface and in lysosomes (particularly secretory lysosomes that contain allergic mediators). In this study, we investigated the role of CD63 in MC using a CD63 knockout mouse model. CD63-deficiency did not affect in vivo MC numbers and tissue distribution. Bone marrow–derived MC developed normally in the absence of CD63 protein. However, CD63-deficient bone marrow–derived MC showed a significant decrease in Fc«RI-mediated degranulation, but not PMA/ionomycin-induced degranulation, as shown by Downloaded from b-hexosaminidase release assays. The secretion of TNF-a, which is both released from granules and synthesized de novo upon MC activation, was also decreased. IL-6 secretion and production of the lipid mediator leukotriene C4 were unaffected. There were no ultrastructural differences in granule content and morphology, late endosomal/lysosomal marker expression, Fc«RI-induced global tyrosine phosphorylation, and Akt phosphorylation. Finally, local reconstitution in genetically MC-deficient Kitw/w-v mice was unaffected by the absence of CD63. However, the sites reconstituted with CD63-deficient MC developed significantly atten- uated cutaneous anaphylactic reactions. These findings demonstrate that the absence of CD63 results in a significant decrease of http://www.jimmunol.org/ MC degranulation, which translates into a reduction of acute allergic reactions in vivo, thus identifying CD63 as an important component of allergic inflammation. The Journal of Immunology, 2013, 191: 2871–2878.

ast cells (MC) originate in the bone marrow and mi- serotonin, proteases, IL-4, and also a significant proportion of grate to most connective tissues and epithelia in pro- TNF-a, are preformed and stored in granules from which they are M cesses that involve various integrins and extracellular released upon MC activation in a process called degranulation or matrix proteins (1). Within those peripheral tissues, MC mature regulated exocytosis. Others, such as numerous cytokines, leu- and can be long-lived, constitutional tissue residents, or recruited kotrienes, and PGs, are synthesized de novo and released upon to sites of inflammation. Their best-established role is in IgE- MC activation (4–8). by guest on September 26, 2021 mediated allergic reactions, but they also positively and nega- MC can be activated by various mechanisms, such as Igs, tively regulate multiple aspects of the inflammatory reaction in cytokines, physical factors, and microbial products. MC activation autoimmunity, contact hypersensitivity, cancer response, and host takes place in acute IgE-mediated allergic reactions such as ana- defense directly and indirectly by influencing dendritic cells, Th1, phylaxis. In this case, Ag-specific IgE bound to its high-affinity Th2, regulatory T cells, Th17, endothelial cells, neurons, and receptor FcεRI expressed at the surface of MC is cross linked possibly B cells (2, 3). These effects are mediated by the nu- by multivalent Ags, leading to the activation of multiple signaling merous mediators that MC release upon activation. Two main cascades. Key features are the activation of protein tyrosine kinases release mechanisms operate in MC. Mediators such as histamine, of the src and spleen tyrosine kinase family, multiple downstream adapter molecules, GTPases and serine-threonine kinases, and calcium mobilization, which ultimately results in the release of *Laboratory of Allergy and Immunology, Department of Pathology, Beth Israel † the different mediators (7, 9, 10). Deaconess Medical Center, Boston, MA 02215; Department of Pathology, Massa- ε chusetts General Hospital, Boston, MA 02114; ‡Institute of Biochemistry, University Fc RI-induced MC activation can be positively or negatively of Kiel, 24098 Kiel, Germany; and xCenter for Vascular Biology Research, Beth affected by a variety of mechanisms (11). One involves tetraspa- Israel Deaconess Medical Center, Boston, MA 02215 nins, proteins that are expressed in various membrane compart- Received for publication August 31, 2012. Accepted for publication July 5, 2013. ments and regulate cell morphology, motility, invasion, fusion, and This work was supported by National Institutes of Health Grant RO1 AI41087 (to signaling (12–14). Using mAb that recognize Ags expressed at the J.-P.K.) and grants from the Deutsche Forschungsgemeinschaft (DFG-GRK1459 and surface of MC and inhibit FcεRI-induced MC degranulation, we SPP1580) and the Center of Excellence Inflammation at Interfaces (to B.S. and P.S.). previously identified the tetraspanins CD63 and CD81 as regu- Address correspondence and reprint requests to Dr. Jean-Pierre Kinet or Dr. Paul ε Saftig, Laboratory of Allergy and Immunology, Department of Pathology, Beth Israel lators of Fc RI-induced MC activation (15, 16). Anti-CD63 Deaconess Medical Center, 99 Brookline Avenue, Boston, MA 02215 (J.-P.K.) or inhibits MC degranulation without affecting early signals, such Institute of Biochemistry, University of Kiel, Otto-Hahn Platz 9, 24098 Kiel, Germany as protein tyrosine phosphorylation and calcium mobilization, (P.S.). E-mail addresses: [email protected] (J.-P.K.) or [email protected] kiel.de (P.S.) suggesting that it interferes with exocytosis (16). CD63 is known The online version of this article contains supplemental material. to be expressed in intracellular membranes, such as secretory Abbreviations used in this article: BMMC, bone marrow–derived mast cell; IKK2, lysosomes (SL) including serotonin-containing granules (17–20). IkB kinase 2; LTC4, leukotriene C4; MC, mast cell; PCA, passive cutaneous anaphy- Upon basophil degranulation, CD63 expression at the plasma laxis; PM, plasma membrane; SL, secretory lysosome; SNARE, soluble N-ethyl- membrane (PM) increases and is classically used as a marker of maleimide–sensitive factor-attachment protein receptor; WT, wild-type. basophil activation (21). Recently, an isoform of CD63 has been Copyright Ó 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00 identified as specific form of CD63 expressed at the PM in www.jimmunol.org/cgi/doi/10.4049/jimmunol.1202323 2872 CD63 IS ESSENTIAL FOR MAST CELL DEGRANULATION IN ALLERGY degranulated MC (22). Anti-CD63 also inhibits adhesion of the immunoblotting were performed using 10% polyacrylamide gels. Anti- MC model cell line RBL-2H3 to extracellular matrix proteins, mouse CD63 rabbit antiserum (25) was used at a 1:400 dilution and ε b which could be explained by the known interaction of CD63 with control anti-Fc RI mAb supernatant [JRK1 (32)] pure, followed by incubation with HRP-conjugated anti-rabbit and anti-mouse Abs (Sigma- integrins in PM microdomains or by inhibition of signaling mech- Aldrich). Anti-phosphotyrosine (4G10; EMD Millipore), anti–phospho- anisms common to FcεRI and integrins (16, 23, 24). Akt (anti–phospho-Ser473 Akt; Cell Signaling Technology, Beverly, MA) Recently, CD63-deficient mice were generated. Despite the and anti-Akt blots (Cell Signaling Technology) of cells triggered as de- broad tissue and cell distribution of CD63, these mice display no scribed for degranulation assays were performed as described (in Refs. 16, 33). obvious morphologic or functional abnormalities of their lysosomes (25). However, they exhibit morphologic changes in the Degranulation assays collecting ducts in the kidneys, a disturbed water balance (25), For assessment of MC degranulation, b-hexosaminidase assays were defective endosomal protein sorting during melanogenesis (26), performed. Briefly, 1 3 106/ml BMMC were loaded with 1 mg/ml anti- and leukocyte recruitment (27). DNP IgE (Sigma-Aldrich) in culture medium without IL-3 overnight. To further investigate the role of CD63 in MC in vivo, we took Then, cells were washed three times in Modified Tyrode’s Buffer (135 mM advantage of the availability of these CD63-deficient mice. NaCl, 5 mM KCl, 1 mM MgCl2, 10 mM HEPES, 5.6 mM dextrose, and 0.1% [w/v] BSA) and then triggered for 30 min at 37˚C with 0–100 ng/ml DNP-BSA (Sigma-Aldrich) or 2–10 nM PMA in combination with 0.5–2 Materials and Methods mM ionomycin at 4 3 106 cells/ml. After stimulation, cells were centri- Mice and bone marrow–derived MCs fuged at 4˚C, and supernatants were incubated in citrate buffer (50 mM citric acid, 50 mM sodium citrate [pH 4.5]) with 0.3 mg/ml p-nitrophenyl The generation and characterization of CD63-deficient mice were described N-acetyl-b-D-glucosaminide [Sigma-Aldrich]) for 60 min at 37˚C. Cell Downloaded from previously (25). MC-deficient Kitw/w-v mice were obtained from The pellets were lysed in a Modified Tyrode’s solution with 1% (v/v) Triton X- Jackson Laboratory (Bar Harbor, ME). Mice were housed in a pathogen- 100 and also incubated in citrate buffer. Reactions were terminated in free barrier facility at the Center for Life Sciences (Boston, MA). Studies carbonate buffer (100 mM Na2CO3 and 100 mM NaHCO3), and after 15 were conducted in accordance with the National Institutes of Health Guide min, OD at 405 nM was determined using an ELISA plate reader. Specific for the Care and Use of Laboratory Animals as well as the Institutional release was calculated as percent of total b-hexosaminidase content after Animal Care and Use Committee guidelines of the Animal Research Fa- subtraction of baseline degranulation in untriggered cells. cility at Beth Israel Deaconess Medical Center. Bone marrow–derived MC

(BMMC) were generated from 4–8 wk-old mice by isolating bone marrows Leukotriene C4 production http://www.jimmunol.org/ from femurs and tibias. Cells were then cultured in IMDM (Corning b Cellgro; Mediatech, Manassas, VA) supplemented with 10% FBS, non- BMMC were loaded and triggered similar to -hexosaminidase assays m except that culture medium for overnight loading contained IL-3 and cell essential amino acids (Corning Cellgro), 50 M 2-ME (Sigma-Aldrich, 3 6 St. Louis, MO), and 3 ng/ml IL-3 (PeproTech, Rocky Hill, NJ). After 4 wk concentration during triggering was 0.5 10 /ml. After triggering, leu- of culture, purity of BMMC cultures was assessed by morphology and flow kotriene C4 (LTC4) and PGD2 production was determined using an EIA cytometric analysis of c-Kit expression using an FITC-conjugated anti–c- assay (Cayman Chemical, Ann Arbor, MI) according to the manufacturer’s Kit mAb (BD Biosciences, San Jose, CA). FcεRI expression was deter- instructions. mined by labeling with IgE (Sigma-Aldrich) and an anti-IgE FITC-conjugated TNF-a and IL-6 ELISA Ab (BD Biosciences). Sections from different organs were subject to formalin fixation and paraffin embedding in a routine histopathologic laboratory BMMC were loaded with 0.5 mg/ml anti-DNP IgE overnight in culture by guest on September 26, 2021 according to standard procedures. Sections were stained with 0.1% to- medium without IL-3. Then cells were washed and incubated for 6 h at luidine blue solution (Sigma-Aldrich). 0.25 3 106 cells/ml in culture medium with 0–100 ng/ml DNP-BSA. ELISA was performed using kits (BD Biosciences) according to the Electron microscopy and Giemsa staining of BMMC manufacturer’s instructions. For electron microscopy, the BMMC were fixed in a dilute mixture of Adoptive transfer of BMMC to Kitw/w-v mice and passive aldehydes, washed in 0.1 M sodium cacodylate buffer, and spun through molten agar to form agar blocks containing the cells. The blocks were cutaneous anaphylaxis processed for electron microscopic studies as described using a Philips Transfer of CD63+/+ and CD632/2 BMMC into MC-deficient Kitw/w-v mice CM10 electron microscope (Philips) (28). In addition, Giemsa stains were 3 6 m (3, 34) was performed by intradermal injection of 1 10 BMMC in PBS performed on 1- m sections from agar blocks and analyzed by light mi- plus 10 mg/ml mouse serum albumin (Sigma-Aldrich) each into ear skin. croscopy. After 4 wk, passive cutaneous anaphylaxis (PCA) was performed as de- Indirect immunofluorescence and confocal microscopy analysis scribed previously (34) with some modifications. Briefly, 100 ng anti-DNP IgE (or control buffer) in PBS plus 10 mg/ml mouse serum albumin was Coverslips were coated in 100 mg/ml poly-L-lysine (Sigma-Aldrich) injected intradermally into ear skin. On the next day, PCA was elicited by overnight and washed twice with water for 1 h before use. BMMC were injecting 200 mg DNP-BSA in 0.9% (w/v) NaCl containing 1% (w/v) fixed in 4% (w/v) paraformaldehyde in PBS for 20 min at room temper- Evan’s blue (Sigma-Aldrich) i.v. into tail veins. After 30 min, mice were ature. Immunocytochemical stainings were performed as described pre- sacrificed, and development of anaphylactic reactions was determined by viously (29) with anti-LAMP1 (1D4B; Developmental Studies Hybridoma measuring the amount of leaked Evan’s blue extracted in formamide (three Bank, Iowa City, IA), anti-LAMP2 (ABL93; Developmental Studies Hy- incubations for 2 h each at 60˚C) from 5-mm ear skin punches. The av- bridoma Bank), anti-cathepsin D (30), anti-CD63 (25), and anti–LIMP-2 erage amount of dye extracted from ears of mice that were injected in (31) as primary Abs. Alexa Fluor 488– or 594–conjugated goat anti-rat IgG buffer (without DNP-IgE) was used as a baseline control and subtracted or goat anti-rabbit IgG were used as secondary Abs (MoBiTec, Go¨ttingen, from test values. Additional mice were subject to PCA, followed by use of Germany), respectively. Coverslips were mounted on glass slides in a ear skin for histopathology and toluidine blue staining. A value of .10% medium containing 17% Mowiol 4-88 (EMD Millipore, Billerica, MA), of granules released from the MC cytoplasm was used to determine the 33% glycerol, and 20 mg/ml 1,4-diaza-bicyclo-[2,2,2]-octane in PBS. number of degranulated MC in tissue sections. Nuclei were visualized with DAPI (Sigma-Aldrich) added to the embed- m ding medium at a concentration of 1 g/ml. Photographs of optical sec- Results tions were acquired with an FV1000 confocal laser scanning microscope 2/2 (Olympus, Tokyo, Japan) equipped with a U Plan S Apo 100˚-oil im- CD63 mice show normal MC numbers and tissue mersion objective (numerical aperture 1.40) and Olympus Fluoview distribution Software (3.0a; Olympus). To evaluate the role of CD63 in MC, CD63-deficient mice were Immunoblotting analyzed in detail (25). Many tetraspanins regulate cell motility in BMMC were solubilized in lysis buffer (50 mM Tris [pH 7.4], 150 mM conjunction with integrins (12). Thus, CD63 deletion could affect NaCl, 5 mM EDTA, and 1% [v/v] Triton X-100), followed by centrifugation the migration, homing, and maturation of MC from the bone and resuspension in SDS sample buffer without 2-ME. SDS-PAGE and marrow into tissues (1). To determine whether CD63-deficient The Journal of Immunology 2873 mice show any abnormalities in MC development and tissue dis- Fc«RI-induced degranulation is significantly diminished in tribution, paraffin-embedded sections from skin, stomach, tra- CD632/2 BMMC cheobronchial tree, and heart were prepared from CD63+/+ and 2/2 To assess the role of CD63 in MC functions, we generated mouse CD63 littermates. The sections were stained with toluidine bone marrow–derived MC (BMMC) from CD63+/+ and CD632/2 blue, which leads to metachromatic staining of MC granules, and 6–8-wk-old littermates by culture in the presence of IL-3. BMMC the numbers of MC and their locations within these tissues were developed normally within 3 to 4 wk of culture and showed no compared. Fig. 1 shows representative views of bronchial tissues obvious differences in morphology (Fig. 2A, 2B), and c-Kit and (Fig. 1A, 1B) and the result of MC counts in skin, stomach, tra- FcεRI expression as determined by flow cytometry (data not cheobronchial tree, and heart from three mice from each genotype shown). To confirm CD63 expression in wild-type (WT) BMMC (Fig. 1C). No differences were found in MC numbers and loca- and loss of expression in genetically deficient BMMC, Western tions within these tissues, indicating that CD63 is dispensable for blot analysis was performed using a recently described polyclonal normal MC development and tissue distribution. Ab (25). As seen in Fig. 2C, BMMC lysates from CD63+/+ mice show a 50–60-kDa band representing heterogeneously glycosy- lated CD63, whereas this band is absent in BMMC lysates from CD632/2 mice. As a control, the same membrane was blotted with an anti-FcεRIb Ab. The expression of FcεRIb was in a comparable range in CD63+/+ and CD632/2 BMMC (Fig. 2D). Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 1. CD632/2 mice show normal numbers and tissue distribution FIGURE 2. BMMC from CD632/2 mice show no overt morphological of MC. Staining of MC in bronchial walls of CD63+/+ (A) and CD632/2 differences to BMMC from CD63+/+ mice and lack CD63 protein ex- (B) mice by toluidine blue (original magnification 3200). (C) Tissue MC pression. Giemsa-stained preparations (original magnification 3400) from numbers in skin, stomach, trachea/bronchi, and heart of CD63+/+ and CD63+/+ BMMC (A) and CD632/2 BMMC (B). Lysates from CD63+/+ and CD632/2 littermates (mean 6 SD; three mice per genotype; 10 high- CD632/2 BMMC were blotted with anti-CD63 (C) or anti-FcεRIb as power fields (HPF) counted twice per condition). a control (D). 2874 CD63 IS ESSENTIAL FOR MAST CELL DEGRANULATION IN ALLERGY

Once we had verified the absence of gross abnormalities of be significantly different (1796 6 137 for CD63+/+ [n = 16]; 1909 6 CD63-deficient MC, we assessed their response to activation by 159 for CD632/2 BMMC [n = 15]; p = 0.59). Thus, a granule comparing FcεRI-induced degranulation in CD63+/+ and CD632/2 storage defect in CD632/2 BMMC is unlikely. In summary, CD63 BMMC. Degranulation was measured by the release of b-hexos- expression is required for efficient MC degranulation via IgE and aminidase after cross linking of FcεRI-bound anti–DNP-IgE with FcεRI in cell-based assays. the multivalent Ag DNP-BSA. Fig. 3A shows that FcεRI-induced CD63-deficient BMMC exhibit decreased TNF2a secretion, MC b-hexosaminidase release is significantly decreased in CD63- whereas IL-6 secretion and leukotriene C synthesis remain deficient BMMC compared with WT littermates (p , 0.0001 by 4 intact ANOVA). PMA and the calcium ionophore ionomycin can induce 2/2 degranulation irrespective from FcεRI activation due to the direct The decrease in degranulation observed in CD63 BMMCs activation of protein kinase C and calcium influx. Fig. 3B shows could be due to a defect in granule functions, resulting from the that degranulation induced by PMA and ionomycin was unaf- absence of CD63 in granules. As opposed to preformed mediators fected in CD63-deficient BMMC (whereas FcεRI-induced trig- stored in granules, such as histamine, serotonin, and hexosamin- gering was still significantly affected in two concomitant experiments idase, other MC mediators such as leukotrienes are synthesized de performed in duplicate with 7.4 6 1.46% b-hexosaminidase release novo (3). Given the role of CD63 in granules/lysosomes, we hy- for CD63+/+ BMMC and 0.0 6 0.4% b-hexosaminidase release pothesized that CD63 deficiency would not affect the release of de for CD632/2 BMMC; p = 0.0016). In addition, total granule novo–synthesized mediators. To test this hypothesis, we measured ε content of b-hexosaminidase (arbitrary units measured from Fc RI-induced leukotriene C4 (LTC4) release in BMMC. As shown

ε Downloaded from lysates plus supernatant at 405-nm wavelength) was calculated in Fig. 4, Fc RI aggregation-induced release of LTC4 is not reduced from all b-hexosaminidase assays performed and found not to in CD63-deﬁcient BMMC. Although cytokines such as IL-6 are mostly synthesized upon activation, TNF-a is both synthesized de novo and stored within MC for rapid release through regulated exocytosis during degranulation (4–6, 8). We hypothesized that TNF-a release upon ε

Fc RI aggregation would be decreased, whereas IL-6 release http://www.jimmunol.org/ would not be affected by CD63 deficiency. As shown in Fig. 5, the absence of CD63 led to impaired TNF-a secretion (Fig. 5A), whereas IL-6 secretion (Fig. 5B) was unaffected. CD63-deficient BMMC show no ultrastructural abnormalities, differences in lysosomal marker expression, and Fc«RI- induced tyrosine and Akt phosphorylation As shown in Fig. 2A and 2B, CD63-deficient BMMC showed no

overt light microscopic differences on Giemsa stains. To further by guest on September 26, 2021 investigate possible differences in granule content and morphology, electron microscopic studies were performed (Fig. 6A, 6B). Analysis of both CD63+/+ and CD632/2 BMMC showed immature MC without discernible ultrastructural differences (two independent experiments performed with n = 3 CD63+/+ and n =2 CD632/2 BMMC cultures). In addition, indirect immunoﬂuores- cence for the late endosomal/lysosomal marker proteins LIMP-2, LAMP-1, LAMP-2, and cathepsin D and confocal microscopy was performed and showed no obvious differences in expression or distribution of these markers, whereas CD63 expression was lost in CD63-deﬁcient BMMC (Fig. 6C–E). Finally, we evaluated proximal FcεRI signaling events. Similar to our prior study using

FIGURE 3. CD632/2 BMMC show diminished FcεRI aggregation-induced degranulation and similar PMA/ionomycin-induced degranulation. CD63+/+ and CD632/2 BMMCs were loaded with anti-DNP IgE and triggered with DNP-BSA for FcεRI-induced activation (A) or a combination of PMA and ionomycin (Iono) (B) for 30 min. Supernatants were assayed for b-hexosaminidase with the substrate p-nitrophenyl N-acetyl- b-D-glucosaminide. Results are expressed as mean + SEM of release calculated as percent of total b-hexosaminidase content and after sub- FIGURE 4. FcεRI aggregation-induced LTC4 production is not signiﬁ- traction of baseline degranulation in untriggered cells (n = 10 for CD63+/+ cantly different in CD63+/+ and CD632/2 BMMC. CD63+/+ and CD632/2 and n = 9 for CD632/2 BMMC cultures from littermates in three inde- BMMCs were loaded with anti-DNP IgE and triggered with DNP-BSA (0– +/+ pendent experiments with p , 0.0001 for (A); n = 6 each for CD63 and 100 ng/ml) for 30 min. LTC4 was measured in the supernatants by enzyme CD632/2 BMMC cultures from littermates in four independent experi- immunoassay. Results are expressed as mean + SEM from three inde- ments with p = 0.8298, p = 0.9597, and p = 0.9892 for 2 nM/0.5 mM, 10 pendent experiments performed each with BMMC from one CD63+/+ and nM/0.5 mM, and 10 nM/2 mM PMA/ionomycin, respectively, for (B). one CD632/2 littermate. The Journal of Immunology 2875 Downloaded from

FIGURE 5. FcεRI aggregation-induced TNF-a secretion, but not IL-6 secretion, is reduced in CD632/2 BMMC. CD63+/+ and CD632/2 BMMCs were loaded with anti-DNP IgE and triggered with DNP-BSA (0–100 ng/ ml) for 6 h. Supernatants were assayed for TNF-a (A) and IL-6 (B)by http://www.jimmunol.org/ ELISA against a standard curve. TNF-a: mean 6 SEM after subtraction of background (one experiment representative of three independent experiments performed each with one CD63+/+ mouse and one CD632/2 littermate). IL-6: one experiment representative of two independent experiments performed each with one CD63+/+ mouse and one CD632/2 littermate. anti-CD63–mediated inhibition of degranulation in RBL (16), global FcεRI-induced tyrosine phosphorylation was unaffected in CD63-deficient BMMC (Supplemental Fig. 1A). However, the by guest on September 26, 2021 Gab2/Akt pathway for degranulation does not appear to be af- +/+ 2/2 fected in CD632/2 BMMC, as shown by retained Akt phosphor- FIGURE 6. CD63 and CD63 BMMC do not show discernible ylation (Supplemental Fig. 1B, 1C). ultrastructural differences and late endosomal/lysosomal marker expression. Electron microcopy of CD63+/+ (A) and CD632/2 (B) BMMC shows CD63 contributes significantly to IgE-mediated anaphylactic immature MC with a large nucleus with partially dispersed chromatin, reactions in vivo narrow surface folds, and many membrane-bound cytoplasmic secretory granules, which are only partially filled with a mixture of membranous and To address the role of CD63 in MC-mediated allergic responses dense materials (data shown are representative for two independent in vivo, we took advantage of the congenitally MC-deficient mouse w/w-v experiments performed in two different facilities [Boston, MA, and Kiel, strain Kit . These mice can be reconstituted with MC of var- Germany]). Original magnifications: (A), left panel 37500; (A), right ious genotypes. This model of MC knockin mice allows for se- panel 37080; (B), left panel 37800; and (B), right panel 37650. Indirect lective analysis of MC-specific functions of genes of interest immunofluorescence and confocal microscopy shows absence of CD63 in vivo (3). MC-deficient Kitw/w-v mice were reconstituted intra- expression in CD632/2 BMMC (C), but comparable expression and dis- dermally in one ear with CD632/2 BMMC and in the other ear tribution of late endosomal/lysosomal proteins LIMP-2 and LAMP-1 (D), with the same number of CD63+/+ BMMC. We verified that equal as well as cathepsin D (Cath. D) and LAMP-2 (E). Scale bars, 10 mm. Data numbers of CD63+/+ and CD632/2 MC were present in each ear shown are representative of two independent experiments. after reconstitution (Fig. 7A–C). We then used an established model of IgE- and FcεRI-mediated anaphylaxis, PCA, adapted for use in locally reconstituted Kitw/w-v from the CD632/2 ear and the CD63+/+ ear was calculated. The mice (34). Anaphylaxis was induced by injecting DNP-specific average of this ratio in seven experiments was 0.49 (SD 0.16; IgE in ears reconstituted with either CD63+/+ or CD632/2 confidence interval 0.34–0.64), which was significantly different BMMC, followed by triggering with DNP-BSA injected i.v. along from one (p = 0.0002; t test). In addition, tissue sections of recon- with Evan’s blue. Evan’s blue binds to albumin and stays in the stituted ears prepared after induction of anaphylaxis in a separate vascular system unless it leaks into the tissues when vascular experiment performed without Evan’s blue were analyzed. These permeability is increased, such as in the case of an acute allergic showed more pronounced MC degranulation in the ear reconstituted reaction. The amount of Evan’s blue that can be extracted from ear with CD63+/+ BMMC, as can be seen by the presence of numerous tissues allows for quantification of the extent of MC degranulation toluidine blue–positive granules outside the MC cytoplasm (Fig. 7 in that ear. In ears reconstituted with CD632/2 BMMC, less E). The majority of CD632/2 BMMC showed no prominent Evan’s Blue dye extravasation was observed than in contralateral degranulation (Fig. 7 F). Additional scoring showed that a mean of ears reconstituted with CD63+/+ BMMC (Fig. 7D). For each ex- 71.01% of CD63+/+ BMMC and 26.19% CD632/2 BMMC were periment, the ratio between the amount of Evan’s blue extracted degranulated (n = 2 slides each). 2876 CD63 IS ESSENTIAL FOR MAST CELL DEGRANULATION IN ALLERGY

FIGURE 7. IgE-mediated passive cutaneous anaphylaxis in ears of MC reconstituted Kitw/w-v mice is reduced in ears reconstituted with CD632/2 BMMC compared with ears reconstituted with CD63+/+ BMMC. Kitw/w-v mice were reconstituted by intrader- Downloaded from mal injection in one ear with 1 3 106 CD63+/+ BMMC and with the same number of CD632/2 BMMC in the other ear. (A) MC numbers were counted per 10 high- power ﬁelds (HPF) on toluidine blue stains. Mean 6 SEM from two independent experiments. (B and C) Toluidine blue stains of sections from ears after WT or 2/2

CD63 BMMC reconstitution (original magniﬁca- http://www.jimmunol.org/ tion 3100). (D) Extravasated Evan’s blue was extracted from ears after PCA and measured. Three independent experiments performed with BMMC from a total of seven CD63+/+ mice and seven CD632/2 littermates. In a separate similar experiment performed without Evan’s blue, the ears were harvested and stained with toluidine blue to assess degranulation in the CD63+/+ BMMC reconstituted ear (E) and the CD632/2 BMMC reconstituted ear (F). Original mag- niﬁcation 3100. by guest on September 26, 2021

Discussion of Abs, possibly due to redundant roles in multimolecular com- In this study, we showed that the tetraspanin CD63 is required plexes (35). for optimal IgE-mediated MC degranulation and anaphylaxis. The discrepancy between the decrease in TNF-a secretion and 2/2 Activation of CD63 MCgeneratedexvivoledtoasignifi- the preservation of IL-6 and LTC4 production is most likely due cantly decreased secretion of b-hexosaminidase and TNF-a, their different processing in MC. TNF-a, like mediators such as b whereas production of IL-6 and LTC4 was unaffected. No ob- serotonin, histamine, and -hexosaminidase, is stored in granules vious ultrastructural differences, differences in late endosomal/ and released upon activation by regulated exocytosis (8). A lysosomal marker expression, as well as alterations in FcεRI- smaller amount of TNF-a issynthesizeduponactivationand dependent tyrosine phosphorylation were seen in CD632/2 released immediately (4, 6). In MC, CD63 is expressed in SL, BMMC. However, MC-deficient mice developed significantly less structures that possess characteristics of both lysosomes and severe passive cutaneous anaphylaxis at sites reconstituted with granules (36). SL are the main organelle responsible for degran- CD632/2 BMMC compared with sites reconstituted with CD63+/+ ulation in nonprofessional secretory cells, such as MC (19, 37– BMMC. These strong effects of CD63 deficiency on MC functions 39). CD63 colocalizes with serotonin in RBL cells, an MC model are unusual, because knockout models on other tetraspanins cell line (20). SL in RBL cells are capable of regulated secretion often have fallen short of confirming the strong functional effects upon FcεRI aggregation. Dynamic studies of MC activated by The Journal of Immunology 2877

FcεRI aggregation have shown the migration of CD63-positive molecules required for both adhesion and degranulation. Given granules toward the PM followed by their fusion with the PM our new data in CD632/2 BMMC, the latter possibility appears (17). As mentioned, we did not observe any decrease of total more likely, given the robust inhibition of degranulation in sus- b-hexosaminidase levels in CD632/2 BMMC, making a storage pended cells. Although RBL cells have been used extensively as defect unlikely. In addition, there were no ultrastructural differ- a model of MC, they may not be entirely representative of primary ences in MC morphology and SL marker expression. This sup- MC. For example, primary MC are heterogeneous depending on ports the hypothesis that CD63 deficiency specifically affects the their degree of maturation, their tissue location, and whether they release of mediators that are stored in SL, either directly or by are resident or recruited to the site of an inflammatory reaction. inhibiting upstream signaling mechanisms. They also differ in the presence of signaling molecules. For ex- The molecular details of distal events of MC degranulation such ample, the tyrosine kinase Fyn, a regulator of FcεRI-induced de- as membrane fusion, and their connection with FcεRI signaling, granulation in BMMC (48), is expressed at only very low levels in are now beginning to be understood. These processes are regulated RBL cells. Therefore, differences should be expected depending by soluble N-ethyl-maleimide–sensitive factor-attachment protein on the source of MC used in the respective study. receptors (SNAREs) (5, 7). The major t-SNAREs (located on the The observed decrease in MC degranulation in CD632/2 target membrane) in MC degranulation appear to be syntaxin 4 BMMC and also with anti-CD63 in RBL cells (16) was ∼50%. and SNAP-23 (40–43), and the major v-SNARE appears to be This strong, but partial inhibition may be due to the role of tet- VAMP-8 (42, 44–46). It remains to be elucidated how CD63 raspanins in multimeric complexes, which incorporate other interfaces with this machinery. Given that PMA/ionomycin- members of the tetraspanin family (12, 35), allowing for other induced degranulation, which circumvents FcεRI-induced signal- tetraspanins to compensate for absent or inhibited CD63. This is Downloaded from ing events that lead to activation of membrane fusion, is unaf- also underscored by the observation that the defects in knockout fected in CD632/2 BMMC, CD63 may be required for signaling models of other tetraspanins were generally mild (including the events that feed into membrane fusion processes. Very proximal previously reported mild effects of CD63 in a knockout model), events such as FcεRI-induced global tyrosine phosphorylation and highlighting possible redundancies (25, 35). Candidate tetraspa- Akt phosphorylation (the latter was affected with anti-CD63 in nins for cooperation with CD63 in MC degranulation are CD81,

RBL) were not affected in CD63-deficient BMMC. However, for which inhibition suppresses degranulation (15), and CD9, which http://www.jimmunol.org/ given the different cellular systems used and the different types of is widely expressed on MC (13). CD63 alterations (binding of surface-expressed CD63 by an Ab in MC-mediated diseases such as allergic rhinoconjunctivitis or the prior study versus removal of both surface-expressed and in- anaphylaxis represent a significant health problem. Our observation tracellular granule–expressed CD63 by genetic targeting in the of significantly impaired IgE-mediated allergic reactions, both current study), differences in affected signaling pathways are to be upon deletion of CD63 (as in this study) or with anti-CD63 Abs (as expected. Regardless, the data obtained with both approaches in the prior studies), indicates that targeting CD63 for therapeutic support an important role for CD63 in the degranulation process purposes may represent a promising avenue for therapy and pos- and allergic reactions in vivo. sibly for other inflammatory diseases in which MC have been A major signaling event during FcεRI-induced degranulation implicated to play a pathogenic role. by guest on September 26, 2021 appears to be the activation of the t-SNARE SNAP23 through IkB kinase 2 (IKK2; IKKb) (43). Interestingly, IKK2-deficient BMMC Disclosures a not only show a defect in degranulation, but also in TNF- se- The authors have no financial conflicts of interest. cretion, similar to our observation in CD63-deficient BMMC in this study. In addition, IKK2 is a signaling molecule downstream of Akt, for which activation was suppressed by anti-CD63 in our References previous study (16). Another recently described role for CD63 is 1. Hallgren, J., and M. F. Gurish. 2011. Mast cell progenitor trafficking and maturation. Adv. Exp. Med. Biol. 716: 14–28. recruitment of the calcium sensor synaptotagmin VII to lyso- 2. Brown, M. A., B. A. Sayed, and A. Christy. 2008. Mast cells and the adaptive somes, which may provide another possible link between FcεRI immune response. J. Clin. Immunol. 28: 671–676. signaling and CD63 (47). Further studies will focus on the role of 3. Galli, S. J., and M. Tsai. 2010. Mast cells in allergy and infection: versatile effector and regulatory cells in innate and adaptive immunity. Eur. J. Immunol. CD63 in the mechanisms that regulate degranulation. 40: 1843–1851. In our prior study using an anti-CD63 mAb and the RBL cell line, 4. Beil, W. J., G. R. Login, M. Aoki, L. O. Lunardi, E. S. Morgan, S. J. Galli, and A. M. Dvorak. 1996. Tumor necrosis factor alpha immunoreactivity of rat we observed that anti-CD63 inhibited degranulation only when peritoneal mast cell granules decreases during early secretion induced by com- tested on adherent cells (16). Anti-CD63 also inhibited RBL cell pound 48/80: an ultrastructural immunogold morphometric analysis. Int. Arch. adhesion to extracellular matrix proteins. In our current study, we Allergy Immunol. 109: 383–389. 5. Benhamou, M., and U. Blank. 2010. Stimulus-secretion coupling by high-affinity used BMMC, which do not adhere spontaneously in contrast to IgE receptor: new developments. FEBS Lett. 584: 4941–4948. RBL, which adhere very strongly to numerous surfaces. For ex- 6. Gordon, J. R., and S. J. Galli. 1990. Mast cells as a source of both pre- ample, to detach RBL cells from culture flasks, incubation of formed and immunologically inducible TNF-alpha/cachectin. Nature 346: 274–276. several minutes at 37˚C with trypsin and EDTA is required. 7. Kalesnikoff, J., and S. J. Galli. 2008. New developments in mast cell biology. CD632/2 BMMC showed diminished degranulation when trig- Nat. Immunol. 9: 1215–1223. 8. Kandere-Grzybowska, K., R. Letourneau,D.Kempuraj,J.Donelan,S.Poplawski, gered in suspension. Therefore, the requirement for adhesion to W. Boucher, A. Athanassiou, and T. C. Theoharides. 2003. IL-1 induces vesicular uncover the role of CD63 in RBL cells may be due to a particu- secretion of IL-6 without degranulation from human mast cells. J. Immunol. 171: larity of these cells not shared by BMMC or due to a CD63 Ab- 4830–4836. 9. MacGlashan, D., Jr. 2008. IgE receptor and signal transduction in mast cells and induced effect that is not seen in cells genetically deficient for basophils. Curr. Opin. Immunol. 20: 717–723. CD63. In addition, the prior study also suggested an inhibitory 10. Rivera, J., and A. Olivera. 2008. A current understanding of Fc epsilon RI- effect of anti-CD63 on the Gab2–PI3K–protein kinase Cd pathway dependent mast cell activation. Curr. Allergy Asthma Rep. 8: 14–20. 11. Kraft, S., and J. P. Kinet. 2007. New developments in FcepsilonRI regulation, of degranulation (16). The possible mechanism of suppression in function and inhibition. Nat. Rev. Immunol. 7: 365–378. those studies included anti-CD63–mediated inhibition of adhesion 12. Hemler, M. E. 2005. Tetraspanin functions and associated microdomains. Nat. Rev. Mol. Cell Biol. 6: 801–811. signaling (which would then lead to diminished degranulation as 13. Ko¨berle, M., S. Kaesler, W. Kempf, F. Wo¨lbing, and T. Biedermann. 2012. a downstream effect) or, alternatively, sequestering of signaling Tetraspanins in mast cells. Front. Immunol. 3: 106. 2878 CD63 IS ESSENTIAL FOR MAST CELL DEGRANULATION IN ALLERGY

14. Yań˜ez-Mo´, M., O. Barreiro, M. Gordon-Alonso, M. Sala-Valde´s, and F. Sańchez- 31. Zachos, C., J. Blanz, P. Saftig, and M. Schwake. 2012. A critical histidine res- Madrid. 2009. Tetraspanin-enriched microdomains: a functional unit in cell plasma idue within LIMP-2 mediates pH sensitive binding to its ligand b-glucocere- membranes. Trends Cell Biol. 19: 434–446. brosidase. Traffic 13: 1113–1123. 15. Fleming, T. J., E. Donnadieu, C. H. Song, F. V. Laethem, S. J. Galli, and 32. Rivera, J., J. P. Kinet, J. Kim, C. Pucillo, and H. Metzger. 1988. Studies with J. P. Kinet. 1997. Negative regulation of Fc epsilon RI-mediated degranulation a monoclonal antibody to the beta subunit of the receptor with high affinity for by CD81. J. Exp. Med. 186: 1307–1314. immunoglobulin E. Mol. Immunol. 25: 647–661. 16. Kraft, S., T. Fleming, J. M. Billingsley, S. Y. Lin, M. H. Jouvin, P. Storz, and 33. Storz, P., and A. Toker. 2003. Protein kinase D mediates a stress-induced NF- J. P. Kinet. 2005. Anti-CD63 antibodies suppress IgE-dependent allergic reac- kappaB activation and survival pathway. EMBO J. 22: 109–120. tions in vitro and in vivo. J. Exp. Med. 201: 385–396. 34. Tsai, M., J. Wedemeyer, S. Ganiatsas, S. Y. Tam, L. I. Zon, and S. J. Galli. 2000. 17. Amano, T., T. Furuno, N. Hirashima, N. Ohyama, and M. Nakanishi. 2001. In vivo immunological function of mast cells derived from embryonic stem cells: Dynamics of intracellular granules with CD63-GFP in rat basophilic leukemia an approach for the rapid analysis of even embryonic lethal mutations in adult cells. J. Biochem. 129: 739–744. mice in vivo. Proc. Natl. Acad. Sci. USA 97: 9186–9190. 18. Blott, E. J., and G. M. Griffiths. 2002. Secretory lysosomes. Nat. Rev. Mol. Cell 35. Hemler, M. E. 2008. Targeting of tetraspanin proteins—potential benefits and Biol. 3: 122–131. strategies. Nat. Rev. Drug Discov. 7: 747–758. 19. Dragonetti, A., M. Baldassarre, R. Castino, M. De´moz, A. Luini, R. Buccione, 36. Griffiths, G. 2002. What’s special about secretory lysosomes? Semin. Cell Dev. and C. Isidoro. 2000. The lysosomal protease cathepsin D is efficiently sorted to Biol. 13: 279–284. and secreted from regulated secretory compartments in the rat basophilic/mast 37. Bonifacino, J. S., P. Perez, R. D. Klausner, and I. V. Sandoval. 1986. Study of the cell line RBL. J. Cell Sci. 113: 3289–3298. transit of an integral membrane protein from secretory granules through the 20. Vincent-Schneider, H., C. The´ry, D. Mazzeo, D. Tenza, G. Raposo, and plasma membrane of secreting rat basophilic leukemia cells using a specific C. Bonnerot. 2001. Secretory granules of mast cells accumulate mature and monoclonal antibody. J. Cell Biol. 102: 516–522. immature MHC class II molecules. J. Cell Sci. 114: 323–334. 38. Raposo, G., D. Tenza, S. Mecheri, R. Peronet, C. Bonnerot, and C. Desaymard. 21. Knol, E. F., F. P. Mul, H. Jansen, J. Calafat, and D. Roos. 1991. Monitoring 1997. Accumulation of major histocompatibility complex class II molecules in human basophil activation via CD63 monoclonal antibody 435. J. Allergy Clin. mast cell secretory granules and their release upon degranulation. Mol. Biol. Cell Immunol. 88: 328–338. 8: 2631–2645. 22. Scha¨fer, T., P. Starkl, C. Allard, R. M. Wolf, and T. Schweighoffer. 2010. A 39. Schwartz, L. B., and K. F. Austen. 1980. Enzymes of the mast cell granule. J. granular variant of CD63 is a regulator of repeated human mast cell degranu- Invest. Dermatol. 74: 349–353. Downloaded from lation. Allergy 65: 1242–1255. 40. Guo, Z., C. Turner, and D. Castle. 1998. Relocation of the t-SNARE SNAP-23 23. Kitani, S., E. Berenstein, S. Mergenhagen, P. Tempst, and R. P. Siraganian. 1991. from lamellipodia-like cell surface projections regulates compound exocytosis in A cell surface glycoprotein of rat basophilic leukemia cells close to the high mast cells. Cell 94: 537–548. affinity IgE receptor (Fc epsilon RI). Similarity to human melanoma differen- 41. Hepp, R., N. Puri, A. C. Hohenstein, G. L. Crawford, S. W. Whiteheart, and tiation antigen ME491. J. Biol. Chem. 266: 1903–1909. P. A. Roche. 2005. Phosphorylation of SNAP-23 regulates exocytosis from mast 24. Nishikata, H., C. Oliver, S. E. Mergenhagen, and R. P. Siraganian. 1992. The cells. J. Biol. Chem. 280: 6610–6620. rat mast cell antigen AD1 (homologue to human CD63 or melanoma 42.Paumet,F.,J.LeMao,S.Martin,T.Galli,B.David,U.Blank,andM.Roa.2000.

antigen ME491) is expressed in other cells in culture. J. Immunol. 149: 862– Soluble NSF attachment protein receptors (SNAREs) in RBL-2H3 mast cells: http://www.jimmunol.org/ 870. functional role of syntaxin 4 in exocytosis and identiﬁcation of a vesicle-associated 25. Schro¨der, J., R. Lu¨llmann-Rauch, N. Himmerkus, I. Pleines, B. Nieswandt, membrane protein 8-containing secretory compartment. J. Immunol. 164: 5850–5857. Z. Orinska, F. Koch-Nolte, B. Schro¨der, M. Bleich, and P. Saftig. 2009. Deﬁ- 43. Suzuki, K., and I. M. Verma. 2008. Phosphorylation of SNAP-23 by IkappaB ciency of the tetraspanin CD63 associated with kidney pathology but normal kinase 2 regulates mast cell degranulation. Cell 134: 485–495. lysosomal function. Mol. Cell. Biol. 29: 1083–1094. 44. Puri, N., and P. A. Roche. 2008. Mast cells possess distinct secretory granule 26. van Niel, G., S. Charrin, S. Simoes, M. Romao, L. Rochin, P. Saftig, subsets whose exocytosis is regulated by different SNARE isoforms. Proc. Natl. M. S. Marks, E. Rubinstein, and G. Raposo. 2011. The tetraspanin CD63 reg- Acad. Sci. USA 105: 2580–2585. ulates ESCRT-independent and -dependent endosomal sorting during melano- 45. Sander, L. E., S. P. Frank, S. Bolat, U. Blank, T. Galli, H. Bigalke, S. C. Bischoff, genesis. Dev. Cell 21: 708–721. and A. Lorentz. 2008. Vesicle associated membrane protein (VAMP)-7 and 27. Doyle, E. L., V. Ridger, F. Ferraro, M. Turmaine, P. Saftig, and D. F. Cutler. VAMP-8, but not VAMP-2 or VAMP-3, are required for activation-induced de- 2011. CD63 is an essential cofactor to leukocyte recruitment by endothelial P- granulation of mature human mast cells. Eur. J. Immunol. 38: 855–863.

selectin. Blood 118: 4265–4273. 46. Tiwari, N., C. C. Wang, C. Brochetta, G. Ke, F. Vita, Z. Qi, J. Rivera, by guest on September 26, 2021 28. Dvorak, A. M. 1987. Procedural guide to specimen handling for the ultrastruc- M. R. Soranzo, G. Zabucchi, W. Hong, and U. Blank. 2008. VAMP-8 segregates tural pathology service laboratory. J. Electron Microsc. Tech. 6: 255–301. mast cell-preformed mediator exocytosis from cytokine trafﬁcking pathways. 29. Savalas, L. R., B. Gasnier, M. Damme, T. Lu¨bke, C. Wrocklage, C. Debacker, Blood 111: 3665–3674. A. Je´ze´gou, T. Reinheckel, A. Hasilik, P. Saftig, and B. Schro¨der. 2011. Dis- 47. Flannery, A. R., C. Czibener, and N. W. Andrews. 2010. Palmitoylation- rupted in renal carcinoma 2 (DIRC2), a novel transporter of the lysosomal dependent association with CD63 targets the Ca2+ sensor synaptotagmin VII membrane, is proteolytically processed by cathepsin L. Biochem. J. 439: 113– to lysosomes. J. Cell Biol. 191: 599–613. 128. 48. Parravicini, V., M. Gadina, M. Kovarova, S. Odom, C. Gonzalez-Espinosa, 30. Pohlmann, R., M. W. Boeker, and K. von Figura. 1995. The two mannose 6- Y. Furumoto, S. Saitoh, L. E. Samelson, J. J. O’Shea, and J. Rivera. 2002. Fyn phosphate receptors transport distinct complements of lysosomal proteins. J. kinase initiates complementary signals required for IgE-dependent mast cell Biol. Chem. 270: 27311–27318. degranulation. Nat. Immunol. 3: 741–748.