The alternatively spliced α_EC domain of human fibrinogen-420 is a novel ligand for leukocyte integrins α_Mβ₂ and α_Xβ₂

In addition to its function in the coagulation and hemostatic systems, plasma protein fibrinogen (Fg) also participates in inflammatory responses. This function of Fg depends on its ability to interact with leukocyte receptors. On inflammatory challenge, Fg mediates adhesive and migratory reactions of leukocytes, such as leukocyte attachment to the vessel wall and subsequent transmigration through the endothelium into a subendothelial matrix.1-3 In addition, fibrinogen and fibrin deposited at sites of vascular injury and within tissues4,5 promote accumulation of inflammatory cells.6,7 The accessory role of Fg in inflammation was documented by in vivo studies in which congenital afibrinogenemia in patients or experimental depletion of Fg from the circulation in animals altered the manifestation of inflammatory responses.8-12 

The interaction of leukocytes with Fg is mediated by 2 transmembrane receptors that belong to the integrin gene superfamily: α_Mβ₂ (CD11b/CD18, Mac-1) and α_Xβ₂ (CD11c/CD18, p150,95). Abundantly expressed on monocytes and neutrophils and induced on lymphocytes during their activation, integrin α_Mβ₂ is primarily responsible for leukocyte adhesion to Fg.13Studies in mice deficient in α_Mβ₂demonstrated that fibrinogen-dependent inflammatory reactions were significantly curtailed in these animals.14 Specifically, the mice failed to accumulate phagocytes at the site of implantation of biomaterials, a process that depends entirely on Fg that is spontaneously adsorbed on the implant surfaces.10 The role of α_Xβ₂, which is enriched on macrophages and dendritic cells, in binding of Fg is poorly characterized.15 16 

Plasma Fg is a dimer composed of 2 copies of 3 nonidentical polypeptide chains (Aα, Bβ, and γ) linked by disulfide bonds.17Structurally, the molecule is organized into a central E domain and 2 peripheral D domains (Figure 1A). The COOH-terminal parts of the Bβ and γ chains in the D domain are folded into the globular βC and γC domains. Previous studies found that the binding site for α_Mβ₂ and α_Xβ₂ in Fg resides in γC.18-20 Two sequences in γC, corresponding to γ190 to 202 and γ377 to 395, were suggested as the putative recognition sites for leukocyte β₂ integrins.21,22Analyses of synthetic peptides duplicating γ190 to 202 and γ377 to 395, designated P1 and P2, respectively, showed that the peptides inhibited adhesion of cells expressing α_Mβ₂and α_Xβ₂, directly supported cell adhesion20,22 and migration,23 and promoted accumulation of leukocytes on implanted biomaterials.24 It was also shown that the sequence γ383 to 395 (P2-C) is the major active site in P2.22 The complementary binding site for P1 and P2 peptides in the heterodimeric α_Mβ₂receptor was localized in the I domain, a region of about 200 amino acid residues inserted in the α_Msubunit.22 25 

In the past decade, it was discovered that normal plasma Fg consists of 2 species differentiated by the length of their Aα chains—a more abundant form with a molecular weight of 340 kd (Fg-340) and a minor form with extended α chains (α_E) and hence a higher molecular weight of 420 kd (Fg-420; Figure 1B).26-28 The globular C-terminal domain of α_E (α_EC), which closely resembles βC and γC, consists of 236 amino acid residues that are missing in the common α chains of Fg-340. In vivo, α_EC originates by alternative splicing of the α gene transcript to include the exon VI sequence.29,30 The properties that the presence of α_EC confers on Fg-420, which is structurally identical to Fg-340 in all other respects, are only beginning to be explored. The properties of the 2 Fg species that are related to classic functions of Fg (ie, clotting, cross-linking by factor XIII, and fibrinolysis) are similar, except that degradation of Fg-420 by plasmin releases a stable product containing the α_EC domain in addition to the conventional degradation products X, Y, D, and E.31 However, little is known about how the presence of α_EC contributes to recognition of Fg-420 by leukocyte integrins. It is known that α_EC and γC share about 40% of amino acid identity,30 and analyses of x-ray structures revealed that folding of the α_EC domain closely resembles that of γC.32-34 In this study, we analyzed the interaction of α_EC with leukocyte integrins. We found that the α_EC domain of Fg-420 is a ligand for α_Mβ₂ and α_Xβ₂capable of mediating strong leukocyte adhesion and promoting directed migration of leukocytes.

Fraction I-2 Fg was purified from umbilical cord plasma, and separation of human Fg-340 and Fg-420 was achieved by ion-exchange chromatography on Mono Q as described previously.31 Fragment D₁₀₀ (M_r100 000) was prepared by digestion of human Fg with plasmin. Fg and plasmin were obtained from Enzyme Research Laboratories (South Bend, IN). Fragment D₁₀₀ was purified by ion-exchange chromatography on CM-Sephadex followed by gel filtration on Sephacryl S-200.35 A recombinant wild-type α_EC domain (α_EC 610-847) was expressed in Pichia pastorisand purified as described previously.36 A recombinant wild-type γC domain was expressed in P pastoris by using an Invitrogen protocol (San Diego, CA). Briefly, a complementary DNA (cDNA) fragment corresponding to γ143 to 411 was generated by polymerase chain reaction from a template consisting of full-length cDNA encoding the human Fg γ chain that was provided by Dr S. Lord (University of North Carolina).37 The amplified cDNA fragment was ligated into pPIC9 expression vector by means ofSnabI and NotI sites and cloned in DH5α cells. The plasmid was linearized by using SalI and transformed into P pastoris (strain GS115) by electroporation. Expression was induced by transferring the yeast into medium containing 0.5% methanol. The recombinant γC was purified by chromatography on Reactive Red-120 CL-6B agarose (Sigma-Aldrich, St Louis, MO). The yeast supernatant was applied to the column, equilibrated with 20 mM Tris-HCl buffer (pH 7.4) containing 80 mM NaCl. Bound protein was eluted with 50 mM Tris-HCl buffer (pH 7.4) containing 0.5 M NaCl. The protein was dialyzed against 50 mM Tris-HCl (pH 7.4) with 0.15 M NaCl, and 45% glycerol was added to prevent aggregation. The purified γC was homogeneous, as revealed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and contained an intact COOH terminus, as demonstrated by Western blot analysis using mAb 4A5, which recognizes γ406 to 411.38 The recombinant α_EC and γC domains were labeled with iodine 125 (¹²⁵I) by using Iodobeads according to the manufacturer's protocol (Pierce, Rockford, IL).

The recombinant α_M I domain corresponding to the human α_M sequence D132 to A318 and the α_X I domain (residues E148-A335) were expressed in Escherichia coli as fusion proteins with glutathione S transferase (GST) by using the expression vector pGEX-4T-1 (Amersham Pharmacia Biotech, Piscataway, NJ) as described previously.39 The GST–I domain fusion proteins from the cell lysate were purified by affinity chromatography using glutathione-agarose (Sigma-Aldrich). To obtain the α_M I domain, the fusion partner was removed by digestion with thrombin and the mixture of GST and α_M I domain was separated by cycling through glutathione-agarose. The Fg peptides P1 and P2-C corresponding to the γ-chain sequences 190 to 202 (Gly-Trp-Thr-Val-Phe-Gln-Lys-Arg-Leu-Asp-Gly-Ser-Val) and 383 to 395 (Thr-Met-Lys-Ile-Ile-Pro-Phe-Asn-Arg-Leu-Thr-Ile-Gly), respectively, were described previously.22 

The mAb 4A5 directed against γ406 to 41138 was provided by Dr G. Matsueda (Bristol-Meyers Squibb, Princeton, NJ). The monoclonal antibodies (mAbs) 44a40 and IB441directed against the α_M and β₂ subunits, respectively, and mAb w6/32 (antihuman HLA) were obtained from the American Type Culture Collection (ATCC; Rockville, MD). Serotec (Raleigh, NC) provided mAb 3.9 directed against the α_Xsubunit. Neutrophil inhibitory factor (NIF) was a gift from Corvas International (San Diego, CA).

Human embryonic kidney (HEK) 293 cells expressing α_Mβ₂ were described previously.42 Chinese hamster ovary (CHO) cells expressing α_Xβ₂ were provided by Dr D. Golenbock (Boston University). The cell lines were maintained in Dulbecco modified Eagle medium (DMEM)–F-12 medium (BioWhittaker, Walkersville, MD) or Ham medium (Life Technologies, Rockville, MD) for α_Mβ₂ and α_Xβ₂transfectants, respectively, and supplemented with 10% fetal-calf serum (FCS) and 25 mM HEPES. The U937 and THP-1 monocytoid cells were obtained from the ATCC and cultured in RPMI 1640 medium supplemented with 10% FCS.

Granulocytes were isolated from peripheral blood obtained from consenting volunteers and anticoagulated with acid-citrate-dextrose essentially as described previously.43 The cells in the preparation were 98% granulocytes, of which at least 96% were neutrophils and 2% eosinophils.

The wells of polystyrene microtiter plates (Costar, Cambridge, MA; or Immulon 4HBX, Dynex Technologies, Chantilly, VA) were coated with various concentrations of protein ligands for 3 hours at 37°C or overnight at 4°C. The wells were postcoated with 1% polyvinylpyrrolidone (PVP) for 1 hour at 37°C. The cells expressing α_Mβ₂ and α_Xβ₂were harvested from the flasks with a cell-dissociating buffer (Gibco, Grand Island, NY) and washed twice in Hanks balanced salt solution (HBSS). The U937 cells, THP-1 cells, and isolated neutrophils were harvested by centrifugation. The cells were labeled with CalceinAM (Molecular Probes, Eugene, OR) for 30 minutes at 37°C, washed with HBSS and resuspended in the same medium at a concentration of 2.5 × 10⁵ cells/mL. Aliquots (100 μL) of the labeled cells were added to each well. For inhibition experiments, cells were mixed with either antibodies, function-blocking peptides, or NIF and incubated for 20 minutes at 22°C before they were added to the coated wells. After 25 minutes of incubation at 37°C in a 3% CO₂ humidified atmosphere, the nonadherent cells were removed and the plates were washed once with HBSS. Fluorescence was measured in a fluorescence plate reader (Perseptive Biosystems, Framingham, MA), and the number of adherent cells was determined from a standard curve constructed by using the fluorescence of 100-μL aliquots with a known number of labeled cells.

Chemotactic motility experiments were done under sterile conditions by using transwell chambers with a pore size 8 μm in diameter (Costar; Corning, Corning, NY). Here, 600 μL DMEM–F-12 medium containing either the D₁₀₀ fragment or the α_EC domain was placed in the lower chamber. The α_Mβ₂-expressing cells or U937 cells harvested as described above were resuspended at a concentration of 2 × 10⁶/mL in DMEM–F-12 and RPMI 1640, respectively. Cells (150 μL) with or without inhibitory antibodies were placed in the upper chamber and incubated for 5 to 15 hours at 37°C in a 3% CO₂ humidified atmosphere. Cells on the upper surface of the polycarbonate membrane were then removed by wiping the surface twice with a cotton-tipped applicator. The membranes were fixed with formaldehyde and stained with hematoxylin 7211 (Richard-Allan Scientific, Kalamazoo, MI). The migrated cells on the lower surface of the filter in 6 randomly chosen low-power fields (×20 magnification) were counted and the values obtained were averaged. No migration was observed in chambers with no added proteins.

Twelve-well microtiter strips (Immulon 2HB; Dynex Technologies) were coated with 100 μL of 10 μg/mL recombinant α_M I domain in Tris-buffered saline (TBS) containing 1 mM CaCl₂and 1 mM MgCl₂ overnight at 4°C and postcoated with 1% bovine serum albumin (BSA) for 1 hour at 22°C. Increasing concentrations of ¹²⁵I-labeled recombinant α_EC or γC in TBS and 0.05% Tween 20 containing 1 mM CaCl₂ and 1 mM MgCl₂ were added to the wells, which were incubated for 3 hours at 37°C. After washing with TBS and 0.05% Tween 20, bound radioactivity was measured and the amount of bound proteins was calculated, with correction for nonspecific binding to BSA-coated wells.

To test the interaction of the α_X I domain, 96-well plates (Immulon 4BX; Dynex Technologies) were coated with 50 μg/mL α_EC and γC overnight at 4°C and postcoated with 3% BSA. Different concentrations of the GST–α_X I domain in 20 mM TBS (containing 1 mM MgCl₂, 1 mM CaCl₂, 0.05% Tween 20, and 5% glycerol) were added to the wells and incubated for 3 hours at 22°C. After washing, bound I domain was detected with an anti-GST mAb (Upstate Biotechnology, Lake Placid, NY) at a 1:5000 dilution. After another washing, goat anti–mouse IgG conjugated to alkaline phosphatase was added for 1 hour, and binding of the I domain was measured by reaction with p-nitrophenyl phosphate. The control binding of GST to immobilized α_EC and γC was typically about 5% to 10% that of the α_M I domain. Background binding to BSA was subtracted.

To determine the amounts of recombinant α_EC and γC immobilized on the wells of the microtiter plates, each domain was adsorbed on the plastic for 3 hours at 37°C at a concentration of 10 or 20 μg/mL and then washed with phosphate-buffered saline. The concentration of bound proteins was determined by the bicinchoninic acid method, according to the manufacturer's protocol (Pierce).

The binding site for α_Mβ₂ in Fg-340 was shown previously to reside in γC, a constituent subdomain of the D domain (Figure 1).18,21,22 Because γC shares about 40% amino acid identity with the α_EC domain of Fg-420,30 and the 2 domains are folded into almost identical structures,33 we tested adhesion of α_Mβ₂-expressing cells to recombinant α_EC. As shown in Figure 2A, when recombinant proteins were immobilized on tissue culture–treated polystyrene plates, both proteins supported efficient adhesion of α_Mβ₂-transfected HEK 293 cells. Adhesion depended on the concentration used, and similar numbers of cells adhered to each protein. The concentrations of α_EC and γC required for half-maximal adhesion were 0.6 μM and 2.1 μM, respectively. The appearance of the cells adherent to the 2 substrates was similar (Figure 2B).

The following results supported the idea that integrin α_Mβ₂ on the surface of α_Mβ₂-transfected cells is responsible for recognition of the α_EC domain. First, mock-transfected cells adhered poorly to α_EC (Figure 2A). Second, mAb 44a and mAb IB4 directed to the α_M and β₂subunits, respectively, inhibited adhesion to α_EC (Figure3A and 3B). Both antibodies inhibited the α_Mβ₂-mediated adhesion to both α_EC and γC almost completely (at 2.5 μg/mL mAb 44A and 10 μg/mL mAb IB4). A control mAb (w6/32) against the class I major histocompatibility complex did not produce inhibition (data not shown). Third, specificity of the interaction between α_Mβ₂ and α_EC was confirmed by the finding that NIF, a specific inhibitor of α_Mβ₂,44 45 abolished adhesion to α_EC completely (data not shown).

We found previously that P1 and P2-C peptides from γC blocked adhesion of α_Mβ₂-bearing cells to the D₁₀₀ fragment, with P2-C being a stronger inhibitor of adhesion.22 Because P1 (γ190-202) and P2-C (γ383-395) sequences are similar to corresponding regions 652 to 664 and 838 to 847 in α_EC, we tested the effect of the P1 and P2-C peptides on adhesion of α_Mβ₂-transfected cells to immobilized α_EC. As shown in Figure4, both P1 and P2-C blocked adhesion to α_EC and D₁₀₀ in a dose-dependent manner. On a molar basis, P2-C was more effective than P1 peptide. For D₁₀₀, 50% inhibition (IC₅₀) was attained with 210 μM P2-C and 350 μM P1. The peptides were less potent inhibitors of adhesion of the cells to α_EC than to immobilized D₁₀₀ (IC₅₀, 65 μM for P2-C and 195 μM for P1). The competence of both P1 and P2-C to inhibit cell adhesion to α_EC seems to be consistent with their ability to cross-inhibit each other's activity22; ie, P2 could inhibit adhesion of the α_Mβ₂-expressing cells to immobilized P1 and vice versa. The ability of P1 and P2-C to inhibit cell adhesion to α_EC and γC suggests that these Fg domains may share common molecular or structural determinants required for recognition by α_Mβ₂. Alternatively, the peptides may inhibit cell adhesion to α_EC and γC by a steric or allosteric mechanism.

The second β₂ integrin, α_Xβ₂, can recognize γC in the D domain of fibrinogen.20 To test the capacity of α_EC to support α_Xβ₂-mediated adhesion, we used CHO cells expressing α_Xβ₂. Adhesion of these cells to several fibrinogen derivatives, including D₁₀₀, was characterized previously.20 Although D₁₀₀includes a second homologous subdomain, βC, earlier studies found that βC does not contribute significantly to the adhesion-promoting activity of D₁₀₀ (Merkulov S. and Ugarova T., unpublished data, 2000). As shown in Figure5A, this study found that α_Xβ₂-expressing CHO cells adhered to α_EC in a dose-dependent manner. Cell adhesion to α_EC was similar to that to D₁₀₀ when the coating concentrations used were in the range of 0 to 2 μg/mL. However, at concentrations greater than 2 μg/mL, the cells continued to show dose-dependent adhesion to α_EC but adhesion to D₁₀₀ declined. Wild-type CHO cells adhered poorly to all 3 substances (α_EC, γC, and D₁₀₀ ligands; 5% adhesion; data not shown). The anti-β₂ mAb IB4, anti-α_X mAb 3.9, and NIF produced 80%, 62%, and 62% inhibition, respectively, of the α_Xβ₂-expressing cells (Figure 5B).

Because of the findings showing that α_EC contains the binding site for α_Mβ₂ and α_Xβ₂, we investigated whether the presence of additional recognition sites in Fg-420 confers greater adhesive potency than that observed with Fg-340. Various concentrations of purified Fg-340 and Fg-420 were deposited on plastic and adhesion of α_Mβ₂-transfected cells was tested. As expected, cells adhered to both Fg forms; however, adhesion to Fg-420 was not significantly greater than that to Fg-340 (data not shown). It is not known whether adhesion to Fg-420, with its multiple binding sites, differs from that of Fg-340 with respect to the cellular activities it triggers.

We next studied the ability of α_EC to support adhesion of U937 and THP-1 monocytoid cells. As shown in Figure6A, U937 cells stimulated with phorbol 12-myristate 12-acetate (PMA) recognized and attached to α_EC strongly, and the adhesion was more extensive than that to D₁₀₀. Similar results were obtained for adhesion of THP-1 (data not shown). Adhesion of both cell lines in the absence of PMA to both immobilized ligands was only slightly less than with PMA stimulation. Adhesion of U937 and THP-1 to α_EC was inhibited by anti-α_M mAb 44a and anti-β₂mAb IB4 (Figure 6B). However, in contrast to the results with α_Mβ₂-transfected cells, inhibition was not complete. The maximal inhibition level achieved at a concentration of 20 μg/mL was 58% with mAb 44a and 77% with mAb IB4.

Because U937 and THP-1 cells express both α_Mβ₂ and α_Xβ₂, the effect of mAb 3.9 against the α_X subunit was tested alone and in combination with mAb 44a. We found that mAb 3.9 inhibited adhesion of U937 cells by 48% and that adding mAb 3.9 to mAb 44a resulted in 75% inhibition. Similar results were obtained with THP-1 cells. Thus, an additional interaction of α_EC with other receptors on the surface of U937 and THP-1 cells may account for the incomplete inhibition of adhesion. It is notable that EDTA, a specific inhibitor of integrin-mediated interaction, completely inhibited adhesion of monocytoid cells to immobilized α_EC, whereas mAb w6/32 against HLA did not produce inhibition (data not shown).

To extend our findings obtained with transfected cells and monocytoid cells, we examined the ability of neutrophils, which are known to express high levels of α_Mβ₂, to bind to α_EC. We found that nonstimulated neutrophils isolated from fresh human blood attached readily to immobilized α_EC and D₁₀₀ in a dose-dependent manner (Figure 7A) and that mAb IB4, which recognizes the common β₂ subunit of α_Mβ₂ and α_Xβ₂on the surface of neutrophils, effectively inhibited adhesion (Figure7B). In addition, mAbs against α_M (44a) and α_X (3.9) inhibited neutrophil adhesion to α_EC by 48% and 33%, respectively (data not shown). The combination of both anti-α–subunit mAbs decreased adhesion by about 70%, similar to the mAb IB4 inhibition. Taken together, these results clearly show that α_EC can support strong adhesion of leukocytes that depends to a large extent on integrins α_Mβ₂ and α_Xβ₂.

In the heterodimeric α_Mβ₂ receptor, the I domain, a region of about 200 amino acid residues inserted in the α_M subunit, contributes importantly to recognition of several ligands, including Fg.46 It was shown previously that synthetic peptides P1 and P2, which duplicate recognition sequences in the γC domain, specifically bound to the recombinant I domain.22,25 Also, in earlier studies, we found that the α_X I domain of α_Xβ₂ is responsible for the binding of P2-C.39 Therefore, in this study, we examined whether α_EC can directly interact with the α_M I and α_X I domains. The recombinant α_M I domain was immobilized on the wells of microtiter plates, and the binding of ¹²⁵I-labeled α_EC and γC was measured. To examine the interaction between α_EC and the α_X I domain, binding of the α_X I domain as a fusion protein with GST was assessed. Figure 8 shows that α_EC bound to the α_M I and α_X I domains in a dose-dependent and saturable manner. The interaction of α_EC and γC with the corresponding I domains seemed to be similar in affinity because the molar concentrations of each ligand required for half-maximal binding were about the same.

Fg was shown previously to promote α_Mβ₂-mediated migration of leukocytes in an in vivo animal model,3 and D₁₀₀ and Fg recognition peptides P1 and P2 supported a chemotactic response of leukocytes in vitro.2,23 47 Therefore, we tested the ability of α_EC to mediate migration of α_Mβ₂-transfected HEK 293 cells and monocytoid cells. The α_Mβ₂-transfected cells were allowed to migrate in a transwell toward a gradient of α_EC or D₁₀₀ at 37°C for 15 hours, after which the cells that migrated through and attached to the underside of the membrane were fixed, stained, and counted. Different concentrations of each protein (range, 1-100 μg/mL) placed in the lower chamber of the transwell system supported efficient migration of the cells, and on a molar basis, D₁₀₀ was a somewhat more efficient inducer of migration than was α_EC (Figure9A). The migration toward α_EC depended on α_Mβ₂, since mAb 44a against the α_M subunit and mAb IB4 against the β₂ subunit blocked the response completely (Figure 9A). Also, α_EC and D₁₀₀ promoted efficient migration of nonstimulated U937 cells (Figure 9B). As with the α_Mβ₂-transfected cells, D₁₀₀was a more potent migratory agent: at a concentration of 50 μg/mL of each protein in the lower chamber, 40 ± 8 cells/field migrated to D₁₀₀, whereas 22 ± 5 cells/field migrated to α_EC. Migration of U937 cells to both ligands was strongly inhibited by mAb 44a, but mAb IB4, the anti-β₂–specific mAb, impeded migration by only 30%. The ability of α_EC to mimic γC in vivo in promoting α_Mβ₂-mediated migration of leukocytes remains to be tested and the advantage of such apparent redundancy explored.

In this study, we showed that leukocytes can bind to α_EC, the domain that distinguishes Fg-420 from Fg-340, by means of the integrins α_Mβ₂ and α_Xβ₂. The recombinant α_EC domain supported adhesion and migration of cells expressing α_Mβ₂ and α_Xβ₂, including cultured monocytoid cells U937 and THP-1 and freshly isolated neutrophils. In addition, cells transfected with cDNA for the 2 integrins bound α_EC. Identification of α_Mβ₂ and α_Xβ₂as receptors for α_EC was based on inhibition of cell adhesion and migration by antibodies against the α_M, α_X, and β₂ subunits; specific interaction of α_EC with the recombinant α_M I domain of α_Mβ₂ and the recombinant α_X I domain of α_Xβ₂; and inhibition of the α_Mβ₂- and α_Xβ₂-mediated adhesion to α_EC by NIF, a specific inhibitor of these integrins. These findings suggest that α_EC could function as an independent recognition site for leukocyte β₂ integrins in Fg-420.

The binding site for α_Mβ₂ and α_Xβ₂ was localized previously to 2 sequences in γC of the D domain of Fg-34020-22: γ190 to 202 (P1) and γ383 to 395 (P2-C).20-22 Although the γC and α_EC domains are structurally very similar,33,34 alignment of amino acid sequences indicates that the P1 and P2-C regions in γC contain only 54% and 31%, respectively, of residues identical to those in corresponding segments of α_EC. Notably, the residues in the C-terminal part of P2-C, (392)Leu-Thr-Ile-Gly(395), responsible for its high cell-adhesive activity22 are not conserved in α_EC. Nevertheless, we found that both γC peptides inhibited adhesion of α_Mβ₂-expressing cells to α_EC, a result consistent with the idea that interaction between α_EC and integrin may involve structural features in common with those of γC.

The data obtained with α_EC and the γC-containing D₁₀₀ fragment indicate that engagement of binding sites on both domains by the α_Mβ₂ integrin on leukocytes can elicit cell migration. Yet the relative availability of those sites in vivo may be a critical differentiating factor permitting initiation of an efficient chemotactic response. During fibrin(ogen)olysis in vitro, the α_EC domains are released early as stable fragments.31 Such α_EC-containing fragments were also detected in plasma obtained from patients with myocardial infarction shortly after initiation of thrombolytic therapy.31 Each fragment is released from Fg-420 by cleavage of a single bond in the proteolytically susceptible region that tethers α_EC to the core of the fibrinogen molecule. Hence, the first degradation products with chemotactic activity released during fibrinolysis would set up a gradient of soluble monomeric α_EC-containing fragments, while γC remained anchored in the D or DD domains incorporated in the fibrin clot. The meaning of these findings may be that α_EC serves as the initial attractant for leukocytes to a thrombus.

The presence in Fg-420 of 2 different interactive regions for leukocyte β₂ integrins, residing in α_EC and γC, raises questions regarding how leukocyte adhesion is regulated and the functional consequences of integrin engagement by these domains together or alone (as in Fg-340, which bears only the γC sites). It is well documented that integrin-mediated cellular adhesion triggers a series of complex signaling events leading to phenotypic changes in cells and that it also initiates gene expression.48Moreover, adhesion to Fg was shown to stimulate synthesis of products relevant to the inflammatory response, including interleukin 1β and oxidative molecules.49,50 Several studies using various integrin-ligand systems demonstrated that differential recognition of various domains in the same or related ligands by a single receptor transduces distinct signals.51-53 For example, occupancy of α₅β₁ on the surface of fibroblasts by different domains of fibronectin induced different cellular responses.52 Fibroblast adhesion to the N-terminal domain of fibronectin induced assembly of novel focal contacts different from those observed after adhesion to the central Arg-Gly-Asp–containing domain. In addition, ligation of α₅β₁ by different fibronectin domains triggered different patterns of tyrosine phosphorylation of signaling molecules. We previously showed that α₄β₁-mediated adhesion of leukocytes to 2 related ligands, vascular cell adhesion molecule 1 (VCAM-1) and the alternatively spliced fragment of fibronectin, IIICS-1, can transduce different signals leading to up-regulation of different sets of matrix metalloproteinases.53 These studies suggest the possibility that engagement of different domains of fibrinogen by means of α_Mβ₂ or α_Xβ₂may also induce distinct patterns of activation of intracellular signaling.

We thank Dr S. E. Plow and L. Zhang for providing the α_Mβ₂-expressing cells, Dr D. Golenbock for providing α_Xβ₂-expressing cells, Dr S. Lord for the γ-chain cDNA, Corvas International for NIF, Dr D. Solovjov for help in preparing the recombinant γC, Peter Baker for assisting with purification of recombinant α_EC, and Timothy Burke for help with neutrophil isolation.