Evidence That Integrin αIIbβ3-dependent Interaction of Mast Cells with Fibrinogen Exacerbates Chronic Inflammation*

Integrin αIIbβ3 is expressed in mast cells as well as in megakaryocytes/platelets. A recent study has shown that surface expression levels of integrin αVβ3 are elevated in integrin αIIb-deficient bone marrow-derived mast cells (BMMCs) as compared with wild-type (WT) counterparts, but the underlying mechanism remains obscure. Here we demonstrate by transducing integrin αIIb into integrin αIIb-deficient BMMCs that surface expression levels of integrin αVβ3 are inversely related to those of integrin αIIbβ3. Thus, competitive association of integrin β3 with integrin αIIb or integrin αV determines surface expression levels of integrin αIIbβ3 or αVβ3 in mast cells. We compared WT and integrin αIIb-deficient BMMCs as well as integrin αIIb-deficient BMMCs transduced with integrin αIIb(WT) or non-functional αIIb(D163A) mutant and found that enhancement of proliferation, degranulation, cytokine production, and migration of BMMCs through interaction with fibrinogen (FB) depended on integrin αIIbβ3. In addition, elevated surface expression of integrin αVβ3 failed to compensate for loss of FB-associated functions in integrin αIIb-deficient BMMCs while enhancing adhesion to vitronectin or von Willebrand factor. Importantly, integrin αIIb deficiency strongly suppressed chronic inflammation with the remarkable increase of mast cells induced by continuous intraperitoneal administration of FB, although it did not affect acute allergic responses or mast cell numbers in tissues in steady states. Interestingly, soluble FB promoted cytokine production of BMMCs in response to Staphylococcus aureus with FB-binding capacity, through integrin αIIbβ3-dependent recognition of this pathogen. Collectively, integrin αIIbβ3 in mast cells plays an important part in FB-associated, chronic inflammation and innate immune responses.

Integrin ␣IIb␤3 is expressed in mast cells as well as in megakaryocytes/platelets. A recent study has shown that surface expression levels of integrin ␣V␤3 are elevated in integrin ␣IIb-deficient bone marrow-derived mast cells (BMMCs) as compared with wild-type (WT) counterparts, but the underlying mechanism remains obscure. Here we demonstrate by transducing integrin ␣IIb into integrin ␣IIb-deficient BMMCs that surface expression levels of integrin ␣V␤3 are inversely related to those of integrin ␣IIb␤3. Thus, competitive association of integrin ␤3 with integrin ␣IIb or integrin ␣V determines surface expression levels of integrin ␣IIb␤3 or ␣V␤3 in mast cells. We compared WT and integrin ␣IIb-deficient BMMCs as well as integrin ␣IIb-deficient BMMCs transduced with integrin ␣IIb(WT) or non-functional ␣IIb(D163A) mutant and found that enhancement of proliferation, degranulation, cytokine production, and migration of BMMCs through interaction with fibrinogen (FB) depended on integrin ␣IIb␤3. In addition, elevated surface expression of integrin ␣V␤3 failed to compensate for loss of FB-associated functions in integrin ␣IIb-deficient BMMCs while enhancing adhesion to vitronectin or von Willebrand factor. Importantly, integrin ␣IIb deficiency strongly suppressed chronic inflammation with the remarkable increase of mast cells induced by continuous intraperitoneal administration of FB, although it did not affect acute allergic responses or mast cell numbers in tissues in steady states. Interestingly, soluble FB promoted cytokine production of BMMCs in response to Staphylococcus aureus with FB-binding capacity, through integrin ␣IIb␤3-dependent recognition of this pathogen. Collectively, integrin ␣IIb␤3 in mast cells plays an important part in FB-associated, chronic inflammation and innate immune responses.
Mast cells play a critical role in IgE-associated allergic disorders, but recent advances have delineated the involvement of mast cells in IgE-independent physiological and pathological processes, including certain innate immune responses. In fact, various stimuli, in addition to IgE and specific antigens, can activate mast cells to release a diverse array of preformed and newly synthesized pro-inflammatory mediators such as histamine, lipids, cytokines, and chemokines (1)(2)(3)(4). Although mast cell numbers and activation in tissues are closely related to mast cell-mediated immunity, the underlying mechanism remains incompletely understood. As one of the key phenomena, mast cells interact with the extracellular matrix (ECM) 2 through integrins composed of two subunits (␣ and ␤), thereby regulating mast cell functions. As previously reported (5)(6)(7)(8)(9), integrins ␣4␤1 and ␣5␤1, or integrin ␣V␤3, expressed in mast cells mediate binding to fibronectin (FN) or vitronectin (VN), respectively. Interestingly, integrin ␣4␤7 is involved in intestinal homing of mast cell progenitors via interaction with mucosal vascular addressin cell adhesion molecule-1 (10). In view of the implication of mast cell integrins in innate immunity, integrin ␣2␤1 expressed in peritoneal mast cells is required for the induction of inflammatory responses to infection (11). In addition, integrin ␣V␤6 is essential for nematode-induced mucosal mast cell hyperplasia and for expression of the granule chymase (12).
Integrin ␣IIb, also known as CD41, which forms a complex with integrin ␤3, is a well known marker of the megakaryocyte/ platelet lineage. Integrin ␣IIb␤3 is required for normal platelet hemostatic function (13)(14)(15)(16)(17). Previously, we reported that integrin ␣IIb␤3 is also highly expressed in mast cells (9,18). In addition, we demonstrated that mast cell interaction with fibrinogen (FB) via integrin ␣IIb␤3 enhances in vitro mast cell functions, by using a blocking Ab specific for integrin ␣IIb (9). On the other hand, higher surface expression levels of integrin ␣V and enhanced adhesion to VN were found in integrin ␣IIbdeficient BMMCs as compared with wild-type (WT) counterparts (18), suggesting that integrin ␣IIb and integrin ␣V counter-regulate their surface expression levels and functions in mast cells. Therefore, we attempted to carefully analyze the regulatory mechanisms by utilizing retroviral transduction with integrin ␣IIb WT or non-functional mutant into integrin ␣IIb-deficient BMMCs.
FB abundant in plasma contributes to blood clotting (19). In addition, FB as well as its degradation product fibrin are also present in ECM outside blood vessels, where they play important roles in inflammation and wound healing through recruitment and activation of inflammatory cells expressing FB-binding receptors (19 -22). Interestingly, surface proteins that bind to FB are also expressed by several types of bacteria, such as Staphylococcus aureus and Streptococcus pyogenes, which modulate immune responses to bacterial infections (23)(24)(25).
In the present study, we showed that integrin ␣IIb expression levels regulate surface expression levels of integrin ␣V␤3 as well as integrin ␣IIb␤3 in mast cells and that mast cell functions augmented by interaction with FB are dependent on integrin ␣IIb␤3 and independent of integrin ␣V␤3. In accordance, integrin ␣IIb deficiency strongly suppressed FB-induced chronic inflammation. Notably, the interaction with soluble FB via integrin ␣IIb␤3 helps mast cells recognize and respond to S. aureus (Cowan I) with FB-binding capacity. Thus, integrin ␣IIb␤3 in mast cells modulates FB-associated, chronic inflammation and innate immune responses.

EXPERIMENTAL PROCEDURES
Mice-All experimental mice were sex-and age-matched (6 -16 weeks old). Balb/c mice were purchased from Charles River Japan (Tokyo, Japan). Integrin aIIb Ϫ/Ϫ mice were generated as described previously (14) and backcrossed to Balb/c mice for at least six generations. Animal studies were performed according to the guidelines of the animal care committee of the Institute of Medical Science, University of Tokyo.
Cells-To generate BMMCs with 95% purity (c-kit ϩ /Fc⑀RI ϩ by flow cytometry), bone marrow cells from 6-week-old male mice were cultured for 5-8 weeks in the presence of 10 ng/ml IL-3 with or without 20 ng/ml SCF as described previously (9,28).
Flow Cytometric Analysis-Cells were stained as described before (9,28). Cells stained with the indicated Abs were analyzed with a FACSCalibur equipped with CellQuest software (BD Biosciences) and Flowjo software (Tree Star).
Adhesion Assay and Migration Assay-Adhesion assay was done as described (6,9). In brief, 96-well plates were coated with 20 g/ml FB, FN, VN, or vWF. BMMCs resuspended at 5 ϫ 10 5 cells/ml were transferred into coated wells with or without stimulant for 1 h at 37°C. After washing, cell adhesion was quantitated using CellTiter-Glo TM (Promega, Madison, WI) and a Micro Lumat Plus luminometer (EG&G Berthold), according to the manufacturer's instructions. In assays using blocking Abs, BMMCs were preincubated with 20 g/ml Abs for 1 h before adding the cells to the plate. Migration assays were carried out as described (7,9), using 24-well Transwell chambers with 5-m polycarbonate filters (Corning).
Measurement of Cytokines-The cells were transferred into FB-coated 96-well plates (1 ϫ 10 4 cells/well) with or without stimulants. After incubating for 12 h at 37°C, the supernatant of each well was collected, and the concentration of IL-6 or TNF-␣ was quantified by enzyme-linked immunosorbent assay with OptiEIA for IL-6 or TNF-␣ (BD Pharmingen) as described (9,28).
PCA Reactions-Passive cutaneous anaphylactic (PCA) reactions were performed as described (32)(33)(34). Briefly, anti-DNP IgE was intradermally injected into the ears of mice. After 24 h, 250 g of DNP-BSA and 0.5% Evans blue dye was intravenously injected. The amounts of extravasated dye were measured after 30 min by extracting ears. In another type of experiment, mice received anti-DNP IgE intravenously. After 24 h, a skin reaction was elicited by applying 0.75% dinitrofluorobenzene acetoneolive oil solution to both sides of the ears. The reaction was assessed by measuring the ear thickness 1 h and 12 h after antigen challenge.
FB-induced Chronic Inflammation Model-200 l of 0.5 mg/ml FB or PBS was intraperitoneally injected into WT or integrin ␣IIb Ϫ/Ϫ mice every 2 days. After 1 month, total peritoneal cells of the sacrificed mice were collected by using 3 ml of PBS. Total cell numbers were counted by using a hemocytometer; the percentages of mast cells (c-kit ϩ /Fc⑀RI ϩ ), granulo-cytes (Gr-1 high /CD11b ϩ ), and macrophages (F4/80 ϩ ) were calculated by fluorescence-activated cell sorting analysis.
Responses to Fixed S. aureus Cowan I in BMMCs-Fixed S. aureus Cowan I (SA) purchased from Calbiochem was stained with Cell Tracker Orange (Molecular Probes). Five ϫ 10 4 BMMCs suspended in 10% BSA/Tyrode buffer were incubated with SA in the presence or absence of 0.5 mg/ml soluble FB in BSA-coated 96-well plates for 2 h. The interaction between mast cells and SA was observed by using a fluorescence microscope. After the supernatant of each well was collected, bacterial cells were removed with a 0.22-m filter. Cytokine concentrations in the supernatants were quantified by enzyme-linked immunosorbent assay.
Quantitation of Tissue Mast Cells-Tissue mast cells in ear skin, back skin, peritoneal wall, and intestine were quantified by light microscopy at ϫ400 by an observer who was unaware of the identity (i.e. mouse genotype) of the individual specimens, in Giemsa-stained sections, as previously described (28,35,36). Results were expressed as mast cells (mean Ϯ S.E.) per mm 2 .
Statistical Analysis-Data are shown as the mean Ϯ S.D. Statistical significance was determined by Student's t test, with p Ͻ 0.01 (**) and p Ͻ 0.05 (*) taken as being statistically significant.

Surface Expression Levels of Integrin ␣V Are Elevated in Integrin ␣IIb-deficient BMMCs-To investigate the role of integrin ␣IIb in mast cells, bone marrow cells from WT and integrin
␣IIb Ϫ/Ϫ mice were cultured in the presence of IL-3 for 5 weeks to generate comparable numbers of morphologically pure (Ͼ95%) mast cells. BMMCs from WT and integrin ␣IIb Ϫ/Ϫ Balb/c mice exhibited similar levels of Fc⑀RI and c-kit on their cell surfaces as determined by flow cytometry (Fig. 1A). In addition, proliferative responses to IL-3 as well as apoptosis induced by growth factor (IL-3) deprivation were comparable between both BMMCs (Fig. 1, C and D). Thus, integrin ␣IIb deficiency did not affect Balb/c mice-derived mast cell development and growth in suspension culture, as previously reported in C57BL/6 mice (18). Moreover, when IgE-sensitized BMMCs were stimulated with the indicated doses of antigen, we found comparable levels of ␤-hexosaminidase release and cytokine (IL-6 and TNF-␣) production ( Fig. 1, E and F, and data not shown). This also suggested that integrin ␣IIb deficiency did not modulate Fc⑀RI signaling in suspension culture of mast cells. However, in keeping with previous findings (18), we confirmed the striking differences between the two cell types: surface expression levels of integrin ␣V and integrin ␣V␤3 were 10-fold higher in integrin ␣IIb-deficient BMMCs as compared with WT counterparts, despite comparable expression levels of other integrins such as integrins ␣4, ␣5, and ␤1, and no detectable expression of integrin ␤2 in either BMMC (Fig. 1E). Collectively, these results led us to postulate that integrin ␣IIb deficiency influenced mast cell functions through interaction with ECM.
Surface Expression Levels of Integrin ␣V Are Inversely Correlated with Those of Integrin ␣IIb-We next investigated the mechanism by which surface expression levels of integrin ␣V were elevated in integrin ␣IIb-deficient BMMCs. As previously reported (18), mRNA levels of integrin ␣V and integrin ␤3 were comparable between WT and integrin ␣IIb-deficient BMMCs (data not shown), suggesting the post-translational regulation of surface expression levels of integrin ␣V in BMMCs. Because integrin ␤3 forms a complex with integrin ␣IIb or integrin ␣V, we hypothesized that integrin ␣V competed with integrin ␣IIb in the association with integrin ␤3. To test this, integrin ␣IIbdeficient BMMCs were retrovirally transduced with integrin ␣IIb WT or mock. Notably, flow cytometric analysis demonstrated that transduction with integrin ␣IIb(WT) strongly down-regulated surface expression of integrin ␣V in integrin ␣IIb-deficient BMMCs ( Fig. 2A). In addition, integrin ␣IIb-(D163A) mutant (37), which lost the capacity to bind to FB, was transduced into integrin ␣IIb-deficient BMMCs. Consistent with a previous report (37), surface expression levels of integrin ␣IIb(D163A) mutant were weaker than those of integrin ␣IIb(WT) in the transduced cells. In proportion to less induction of integrin ␣IIb(D163A), surface expression levels of integrin ␣V in integrin ␣IIb(D163A) mutant-transduced cells were less down-regulated as compared with those in integrin ␣IIb(WT)-transduced BMMCs ( Fig. 2A). Furthermore, similar experiments were performed using murine T cell lymphoma cell line BW5147, which originally expressed integrin ␣V␤3 but not integrin ␣IIb␤3. As shown in Fig. 2B, transduction with integrin ␣IIb(WT) into BW5147 cells down-regulated surface expression of integrin ␣V to a greater degree as compared with transduction with integrin ␣IIb(D163A) mutant. Collectively, surface expression levels of integrin ␣V were inversely related to those of integrin ␣IIb, and even non-functional integrin ␣IIb competed with integrin ␣V for integrin ␤3.
Reduced Adhesion to FB and Enhanced Adhesion to VN and vWF in Integrin ␣IIb-deficient BMMCs-Next, we examined the effects of integrin ␣IIb deficiency on mast-cell adhesion to ECM proteins such as FN, FB, VN, and vWF. IgE stimulationdependent adhesion to FB was strongly suppressed in integrin ␣IIb-deficient BMMCs, whereas adhesion to VN or vWF was drastically enhanced in integrin ␣IIb-deficient BMMCs, presumably because of increased surface expression of integrin ␣V␤3 in integrin ␣IIb-deficient BMMCs when compared with WT BMMCs (Fig. 3A). On the other hand, integrin ␣IIb deficiency did not significantly affect the adhesion to FN (Fig. 3A). In addition, we examined the inhibitory effect of pretreatment with blocking Abs against integrin ␣IIb␤3 or integrin ␣V␤3 on mast-cell adhesion to ECM proteins, confirming that, in WT BMMCs, the binding to FB, VN, or vWF was dependent on integrin ␣IIb␤3, integrin ␣V␤3, or both (Fig.  3B). Similar experiments were also performed with regard to integrin ␣IIb-deficient BMMCs, demonstrating that pretreatment with blocking Abs against integrin ␣V dampened IgE stimulation-dependent strong adhesion to VN or vWF as well as weak adhesion to FB, whereas it did not affect the adhesion to FN (Fig. 3B). In contrast, pretreatment with blocking Ab for integrin ␣IIb did not reduce the adhesive property at all (Fig. 3B). Collectively, these results suggested that elevated surface expression levels of integrin ␣V␤3 in ␣⌱⌱b-deficient BMMCs enhanced the adhesion to VN or vWF, but it did not compensate for the defective adhesion to FB owing to integrin ␣IIb␤3 deficiency.

Enhancement of Migration, Proliferation, Degranulation, and Cytokine Production of BMMCs through Interaction with FB Is
Dependent on Integrin ␣IIb-We next examined the effect of integrin ␣IIb deficiency on mast-cell functions through interaction with FB. As previously reported, SCF induced migration of WT BMMCs when the lower membranes of the Transwells were pre-coated with FB, FN, or VN. Comparison of the migrating cell numbers between WT and integrin ␣IIb-deficient BMMCs revealed that integrin ␣IIb deficiency strongly diminished or enhanced the migration of BMMCs through interaction with FB or VN, respectively, whereas it did not affect mast cell migration through interaction with FN (Fig. 3C). These results also suggested that, in integrin ␣IIb-deficient BMMCs, both the adhesive and migratory ability were altered toward integrin ␣V␤3, whereas there was little interaction with FB, a specific ligand for integrin ␣IIb␤3. Moreover, it was found in WT, but not integrin ␣IIb-deficient, BMMCs that SCF-stimulated mast-cell proliferation was accelerated in FBcoated plates as compared with BSA-coated plates (Fig. 4A). Similarly, when stimulated by IgE plus antigen, WT, but not integrin ␣IIb-deficient, BMMCs enhanced ␤-hexosaminidase release and cytokine (IL-6 and TNF-␣) production through interaction with FB (Fig. 4,  B-D). Altogether, integrin ␣IIb␤3 plays crucial roles in enhancing mast-cell functions through interaction with FB.

Mutant, into Integrin ␣IIb-deficient BMMCs Recovered Mast Cell Functions through Interaction with FB-
To further reduce the possibility that enhanced expression levels of integrin ␣V␤3 modulated mast-cell functions through interaction with FB, we performed similar experiments on adhesion and cytokine production in integrin ␣IIb-deficient BMMCs transduced with integrin ␣IIb(WT), integrin ␣IIb-(D163A) mutant, or mock. As depicted in Fig. 5A, transduction with integrin ␣IIb(WT), but not integrin ␣IIb(D163A) mutant, augmented the adhesion to FB as compared with transduction with mock. On the other hand, transduction with integrin ␣IIb(D163A) as well as integrin ␣IIb(WT) diminished the adhesion to VN, with the degree of the former being a little lower than that of the latter, which was consistent with surface expression levels of integrin ␣V␤3 in integrin ␣IIb-deficient BMMCs transduced with integrin ␣IIb(WT) and (D163A) mutant ( Fig. 2A). Moreover, transduction with integrin ␣IIb(WT), but not integrin ␣IIb(D163A), induced the enhancement of cytokine production through interaction with FB in integrin ␣IIb-deficient BMMCs (Fig. 5B). Collectively, these results definitively confirmed that enhanced mast cell functions through interaction with FB were dependent on integrin ␣IIb␤3 but not integrin ␣V␤3.
Integrin ␣IIb Deficiency Affected Neither Tissue Mast Cell Numbers in Steady States nor Mast Cell-mediated Acute Allergic Reactions-Because enhanced proliferation and migration of BMMCs through interaction with FB were suppressed by integrin ␣IIb deficiency, we compared the quantity of tissue mast cells in WT and in integrin ␣IIb Ϫ/Ϫ mice. Microscopic analysis demonstrated that mast-cell numbers in the ear skin, back skin, peritoneum wall, and small intestine were not different in these mice (Table 1). Based on this, we addressed the question of whether tissue FB extravasated by acute inflammation modulated mast cell-associated allergic reactions of WT and integrin ␣IIb Ϫ/Ϫ mice. However, no significant difference of two types of PCA reaction was observed in these mice (data not shown), despite enhanced in vitro degranulation and cytokine production of mast cells through integrin ␣IIb␤3-depend-ent interaction with FB (Fig. 4,  B-D). These results indicated that integrin ␣IIb␤3 was not involved in tissue mast-cell numbers and distributions in steady states or IgE-mediated acute allergic responses.
Integrin ␣IIb Deficiency Suppressed Peritoneal Chronic Inflammation with a Remarkable Increase of Mast Cells Induced by Repetitive Intraperitoneal FB Administration-We next asked whether integrin ␣IIb deficiency influenced chronic inflammation with extravascular FB and fibrin deposition. To explore the direct effects of FB, we adopted FB-induced chronic inflammation models where FB was administered into peritoneal cavities every other day. After 1 month, we counted total peritoneal cell numbers and estimated cell populations by flow cytometric analysis. In steady states before the stimulation, we found no significant differences in total peritoneal cell numbers or in mast cell numbers between WT and integrin ␣IIb Ϫ/Ϫ mice ( Fig. 6A and data not  shown). Interestingly, repetitive intraperitoneal injection of FB, but not PBS as a control, induced severe chronic inflammation with a remarkable increase of mast cells as well as total inflammatory cells in the peritoneal cavities of WT mice (Fig. 6, A and B). Thus, an FB-induced chronic inflammation model was established. Intriguingly, integrin ␣IIb deficiency strongly suppressed the number of mast cells as well as the total number of inflammatory cells in the peritoneal cavities (Fig. 6, A and B), although the percentages of granulocytes and macrophages were not significantly different in the WT and integrin ␣IIb Ϫ/Ϫ mice (Fig. 6B, right panel). Considering the in vitro roles of mast cell integrin ␣II␤3-FB interaction, these results strongly suggested that in vivo FB-induced chronic inflammation was largely dependent on integrin ␣IIb␤3 in mast cells, although the effect of few, if any, platelets in the peritoneal cavities on this phenomenon was not completely ruled out. In addition, we found comparable numbers of inflammatory cells in WT and integrin ␣IIb-deficient mice 24 h after single dose of FB injection (data not shown). Therefore, FB-induced chronic inflammation required continuous administration of FB. Collectively, integrin ␣IIb␤3 in mast cells played an important role in FB-mediated chronic, but not acute, inflammatory responses.
Soluble FB Enhanced Cytokine Production of WT, but Not Integrin ␣IIb-deficient BMMCs, in Response to S. aureus Cowan I with FB-binding Capacity-As previously reported, mast cells adhered to soluble FB as well as plate-coated FB via integrin ␣IIb␤3. Because soluble FB is bound by certain types of bacteria such as S. aureus (Cowan I), the immune cells expressing FB-binding receptors are thought to modulate the immune responses to these pathogens (23)(24)(25). We then investigated whether soluble FB influenced the response of mast cells to S. aureus (Cowan I). When WT or integrin ␣IIb-deficient BMMCs were incubated with S. aureus (Cowan I) for 2 h in the presence of soluble FB, fluorescent microscopic analysis demonstrated that WT, but not integrin ␣IIb-deficient, BMMCs were completely surrounded by aggregated S. aureus (Cowan I) probably through interaction with soluble FB (Fig. 7A). On the other hand, BMMCs were not apparently covered with S. aureus (Cowan I) in the absence of soluble FB. These results suggested that integrin ␣IIb␤3-dependent interaction of BMMCs with S. aureus (Cowan I) via soluble FB probably helped mast cells recognize this pathogen. Moreover, IL-6 released into each supernatant was quantified by enzymelinked immunosorbent assay, demonstrating that soluble FBinduced enhancement of IL-6 production was observed only in WT, but not integrin ␣IIb-deficient, BMMCs in response to S. aureus (Cowan I) (Fig. 7B). To examine the specificity of this phenomenon, similar experiments were performed using Escherichia coli without FB-binding capacity. As shown in Fig. 7B, soluble FB-dependent enhancement of IL-6 production of BMMCs stimulated by E. coli was not observed irrespective of integrin ␣IIb expression, suggesting that soluble FB induced the enhancement of cytokine production of BMMCs in response to bacteria with, but not without, FB-binding capacity. Because Toll-like receptors primarily play an important part in the recognition of and response to bacteria (3), we also asked if soluble FB or immobilized FB enhanced cytokine production of both BMMCs stimulated by LPS, a Toll-like receptor 4 agonist. As    Fig. 7 (B and C), soluble FB did not affect IL-6 production of either WT or integrin ␣IIb-deficient BMMCs stimulated by LPS, whereas immobilized FB enhanced IL-6 production of WT, but not integrin ␣IIb-deficient, BMMCs stimulated by LPS. These results suggested the synergism of Toll-like receptor 4 signaling and integrin ␣IIb␤3 signaling through interaction with immobilized FB, but not soluble FB. Altogether, soluble FB enhances the cytokine production of BMMCs in responses to S. aureus (Cowan I), probably because mast cell-soluble FB-S. aureus (Cowan I) complex formation promoted the quick and tight recognition of this pathogen by mast cells.

DISCUSSION
In a previous study, we found that integrin ␣IIb␤3 is highly expressed in mast cells, in addition to the megakaryocyte/platelet lineage and a subset of hematopoietic progenitors (9,14,18). Experiments using blocking Abs specific for integrins demonstrated that adhesion to FB, VN, or vWF was mediated through integrin ␣IIb␤3, integrin ␣V␤3, or both, respectively (9). In the follow-up study, we first paid attention to the interesting results shown by Berlanga O et al. that integrin ␣IIb-deficient BMMCs displayed extremely higher surface expression levels of integrin ␣V␤3 as compared with WT counterparts (18). Because counter-regulation of integrin ␣IIb␤3 and integrin ␣V␤3 on their surface expression levels might affect in vivo functions of integrin ␣IIb␤3 in mast cells, we attempted to delineate the underlying mechanism. Our hypothesis that integrin ␣IIb competed with integrin ␣V in heterodimerization with integrin ␤3 in mast cells was illustrated by experimental results as follows: retroviral transduction with integrin ␣IIb(WT) into integrin ␣IIb-deficient BMMCs reduced surface expression of integrin ␣V␤3 at levels comparable to those in WT BMMCs ( Fig. 2A). In addition, transduction with integrin ␣IIb(D163A) mutant led to less reduction in surface expression levels of integrin ␣V␤3 together with less induction in those of integrin ␣IIb␤3 in integrin ␣IIb-deficient BMMCs. Thus, surface expression levels of integrin ␣V␤3 were conversely related to those of integrin ␣IIb␤3 in mast cells ( Fig. 2A). Notably, this phenomena was true for BW5147 cells transduced with integrin ␣IIb(WT) or  . Soluble FB enhanced cytokine production of WT, but not integrin ␣IIb-deficient, BMMCs in response to S. aureus with FB-binding capacity. A, WT or integrin ␣IIb-deficient BMMCs were incubated with heatkilled S. aureus labeled by Cell Tracker Orange in the presence or absence of 500 g/ml soluble FB for 2 h. WT, but not integrin ␣IIb-deficient, BMMCs were covered with SA aggregates in the presence of soluble FB (arrowhead). B, WT or integrin ␣IIb-deficient BMMCs were incubated with 100 ng/ml LPS, 100 g/ml heat-killed E. coli, 100 g/ml S. aureus, or PBS as control in the presence or absence of 500 g/ml soluble FB for 8 h. The ratio of the amounts of IL-6 released in the presence of soluble FB to those of IL-6 in the absence of soluble FB was measured. Data are representative of three independent experiments. All data points correspond to the mean Ϯ S.D. ** (p Ͻ 0.01) indicates statistical differences. C, WT or integrin ␣IIb-deficient BMMCs were incubated with the indicated concentrations of LPS in FB-or BSA-coated plates. Data represent three independent experiments. All data points correspond to the mean Ϯ S.D. * (p Ͻ 0.05) indicate statistical differences. ␣IIb(D163A) mutant (Fig. 2B). However, integrin ␣IIb deficiency did not affect surface expression levels of integrin ␣V␤3 in platelets (data not shown). These results suggested that regulatory mechanisms on surface expression levels of integrin ␣V␤3 differed between mast cells and platelets. One possible explanation is as follows: integrin ␣IIb deficiency might fail to influence surface expression levels of integrin ␣V␤3 if integrin ␤3 expression were sufficient in platelets, whereas it might promote the association of integrin ␣V with integrin ␤3 if integrin ␤3 expression were insufficient in mast cells. Further examination is necessary to fully understand the mechanism. Importantly, all the functional analyses (Figs. 2-4) showed that higher surface expression levels of integrin ␣V␤3 in integrin ␣IIb-deficient BMMCs enhanced adhesion to VN or vWF but did not compensate for the loss of mast-cell functions through interaction with FB. Based on this, we compared in vivo mast cell functions between WT and integrin ␣IIb-deficient mice.
First, integrin ␣IIb deficiency did not affect mast-cell numbers in tissues under normal conditions (Table 1). This seems reasonable, given that FB is not abundant outside blood vessels under normal conditions. Second, integrin ␣IIb deficiency did not affect two types of PCA estimated by ear dye extravasation or swelling (data not shown), which was reported to be mast cell-dependent (32)(33)(34). These results indicate that extravascular FB and fibrin accompanied by acute inflammation are unable to enhance mast-cell functions. On the other hand, recent advances demonstrate that FB is a central regulator of the inflammatory response as well as of hemostasis. Analysis of gene-targeted mice expressing a mutant form of FB, lacking the integrin ␣ M ␤2-binding motif, demonstrated that the high affinity engagement of FB by integrin ␣ M ␤2 in neutrophils and macrophages was critical for inflammatory responses (38,39). Therefore, we speculated that FB extravasated at acute inflammatory sites activated neutrophils and macrophages via integrin ␣ M ␤2 but failed to enhance mast-cell functions via integrin ␣IIb␤3. In contrast, we found striking differences between WT and integrin ␣IIb Ϫ/Ϫ mice in FB-induced chronic inflammation: integrin ␣IIb deficiency strongly suppressed the increase of total inflammatory cells with mastocytosis in the peritoneal cavities. However, administration of single dose FB did not lead to any significant difference of initial inflammatory responses in these mice 24 h after inoculation (data not shown), confirming the negligible role of integrin ␣IIb in acute inflammation. Taking into consideration that platelets are absent in the peritoneal cavities and that integrin ␣V␤3 did not significantly affect in vitro mast-cell functions through interaction with FB, we concluded that FB-induced chronic inflammation depended on integrin ␣IIb␤3 in mast cells. The relevant mechanism might be as follows: FB activates macrophages and granulocytes via integrin ␣ M ␤2 to produce inflammatory cytokines and chemokines in the initial phase, leading to the gradual recruitment, proliferation, and activation of mast cells in the presence of FB. Alternatively, activated mast cells also produce a diverse array of chemical mediators, accelerating chronic inflammation. Thus, FB-mediated inflammation appears to be augmented with the increase of mast cells in tissues. This scenario may explain in part why mast cell numbers increase in a variety of chronic inflammatory diseases such as atopic derma-titis and asthma that are thought to cause continuous extravasation of FB in tissues. Further analysis of WT and integrin ␣IIb-deficient mice using different types of chronic inflammation models will be required to delineate the role of mast cell integrin ␣IIb␤3 in chronic inflammatory diseases.
Another important finding in this study was that soluble FB enhanced IL-6 production of WT, but not integrin ␣IIb-deficient, BMMCs in response to S. aureus (Cowan I) with FBbinding capacity. On the other hand, soluble FB failed to enhance IL-6 production of WT BMMCs stimulated by E. coli harboring no FB-binding capacity, LPS, or bacterial lipopeptide ( Fig. 7B and data not shown). Because WT, but not integrin ␣IIb-deficient, BMMCs apparently kept a strong contact with aggregated S. aureus (Cowan I) in the presence of soluble FB, integrin ␣IIb␤3-dependent recognition of S. aureus (Cowan I) in mast cells may augment the innate response to this pathogen. Considering that soluble FB facilitates the interaction of platelets with S. aureus by bridging clumping factor A in S. aureus and integrin ␣IIb␤3 in platelets (23)(24)(25), a similar mechanism probably occurs in mast cells: the complex formation of mast cell integrin ␣IIb␤3-FB-S. aureus (Cowan I) promotes quick and tight recognition of this pathogen by mast cells, thereby enhancing innate immune responses. Collectively, these results suggested that integrin ␣IIb␤3 plays an important part in the innate responses of mast cells to certain types of bacteria with FB-binding capacity.
In conclusion, the integrin ␣IIb␤3-dependent interaction of mast cells with FB augments FB-associated chronic inflammation or innate responses to FB-binding bacteria. Elucidation of the in vivo function of integrin ␣IIb␤3 in mast cells will lead to new approaches in the prevention of and therapy for the relevant inflammatory and infectious diseases.