Molecular Identification of the Cross-reacting Epitope on αMβ2 Integrin I Domain Recognized by Anti-αIIbβ3 Monoclonal Antibody 7E3 and Its Involvement in Leukocyte Adherence*

The monoclonal antibody (mAb) 7E3 directed to the platelet integrin αIIbβ3 was tested for its cross-reactivity with the homologous leukocyte integrin αMβ2. Nested recombinant fragments of αM I domain were expressed as glutathioneS-transferase fusion proteins and analyzed for antibody recognition. In enzyme-linked immunosorbent assay, mAb 7E3 bound αM I domain fragments containing the amino-terminal sequence Cys128–Ser172, whereas the carboxyl-terminal region Leu173–Pro291 was ineffective. A synthetic peptide designated R1.1 and duplicating the αM sequence G127CPQEDSDIAFLIDGSGSIIPHDF150 bound mAb 7E3. In contrast, the adjacent αM region F150RRMKEFVSTVMEQLKKSKTLFS172 or a control peptide with a scrambled R1.1 sequence was not recognized by mAb 7E3. Binding of mAb 7E3 to αM I domain blocked monocyte and neutrophil adhesion to immobilized fibrinogen and fibrinogen-dependent leukocyte-endothelium bridging, indistinguishably from bona fide anti-β2 mAb IB4. In contrast, leukocyte binding to stable transfectants expressing intercellular adhesion molecule-1 was not affected by mAb 7E3. Balloon-mediated injury of iliofemoral arteries in rabbits resulted in prominent deposition of fibrinogen and increased monocyte adhesion to the injured vessel, in a reaction inhibited by mAb 7E3, but unaffected by control mAb 14E11. Through its cross-reactivity between αIIbβ3 and αMβ2, mAb 7E3 may initiate a new class of integrin antagonists, capable of simultaneously targeting platelet and leukocyte adhesion mechanisms in vascular injury.

Hemostasis and immune-inflammatory responses (1) are maintained by the adhesive interactions mediated by integrins ␣ IIb ␤ 3 (GPIIb/IIIa) on platelets (2) and ␣ M ␤ 2 (Mac-1) on leukocytes (3). Despite their critical role in vascular cell homeostasis and signaling (4), platelet and leukocyte adhesion mechanisms participate in the pathogenesis of vascular injury. This is emphasized by the role of ␣ IIb ␤ 3 in platelet aggregation and thrombus formation (5,6) and of ␣ M ␤ 2 in leukocyte recruitment (3), procoagulant activity (7), and reperfusion injury. Consid-erable effort has been devoted to the identification of molecular antagonists of ␣ IIb ␤ 3 (9) and ␣ M ␤ 2 (10), capable of disrupting aberrant platelet and monocyte adherence mechanisms. In this context, administration of anti-␣ IIb ␤ 3 mAb 1 7E3 (11) reduced the incidence of mortality, myocardial infarction, and other emergency procedures in patients at risk of cardiovascular ischemic disease (12,13).
In previous studies, it was also reported that mAb 7E3 unexpectedly cross-reacted with the active conformation of ␣ M ␤ 2 , induced on monocytes by inflammatory stimuli (14) or Mn 2ϩ ions (15). These observations were recently independently confirmed with direct binding studies of mAb 7E3 to ␣ M ␤ 2 transfectants (16), whereas a ϳ200-amino acid-inserted "I" domain in ␣ M (17) was provisionally implicated in this cross-reactivity (18).
In this study, we sought to reinvestigate the molecular basis of mAb 7E3 cross-reactivity with ␣ M ␤ 2 and its potential relevance to leukocyte adhesion, in vivo. We found that mAb 7E3 recognizes a discrete region in ␣ M I domain (17), which is critically involved in monocyte adherence to fibrinogen in vitro and in balloon-injured arteries, ex vivo.
Recombinant ␣ M I Domain Fragments-The map of the various ␣ M I domain fragments used in these studies is shown in Fig. 1. For these experiments, the full-length ␣ M cDNA was amplified by PCR in the presence of a forward oligonucleotide 5Ј-TGTCCTCAAGAGGATAGT-GAC-3Ј and four distinct reverse oligonucleotides 5Ј-AGAGAACAAG-GTTTTGGAC-3Ј (R1), 5Ј-CGTGGCCGTGTGTGTC-3Ј (R2), 5Ј-CT-CATATCCCAAGGGATCG-3Ј (R3), and 5Ј-CGGCTTGGATGCGATG-3Ј (R4). The fragment R1(Ϫ), lacking the amino-terminal R1 sequence Cys 128 -Ser 172 (Fig. 1), was generated by PCR using forward and reverse oligonucleotides 5Ј-TTGATGCAGTACTCTGAAG-3Ј and 5Ј-CGGCTTG-GATGCGATG-3Ј, respectively. The ␣ M I domain fragments B5 and B4 ( Fig. 1) were generated with forward and reverse primers 5Ј-TTGATG-CAGTACTCTGAAG-3Ј and 5Ј-ATTCTTTCGGGCTCCGTTG-3Ј (B5) and 5Ј-GCCTTTAAGATCCTAGTTGTC-3Ј and 5Ј-CGGCTTGGATGC-GATG-3Ј (B4), respectively. Each forward primer contained a BamHI restriction site, whereas a XhoI site was added at the end of each reverse oligonucleotide. Amplification was carried out in a total volume of 100 l with denaturation at 94°C for 1 min, annealing at 52°C for 1 min, and extension at 72°C for 1 min. PCR products were separated on 1% agarose gels, gel-purified by phenol/chloroform extraction, digested overnight with BamHI and XhoI, and directionally cloned in the prokaryotic expression vector pGEX-2T (Amersham Pharmacia Biotech) with transformation in the BL-21 Escherichia coli strain. Expression of recombinant proteins was carried out as described (18). Briefly, E. coli cultures containing the various constructs were grown to A 600 ϳ0.5, induced with 0.1 mM isopropyl-␤-D-thiogalactopyranoside (IPTG, Calbiochem) and grown for 3 h at 37°C with constant shaking at 225 rpm. Bacteria were centrifuged at 6,000 rpm for 15 min at 4°C and suspended in a lysis buffer containing 10 mM Tris-HCl, 150 mM NaCl, 100 g/ml lysozyme (Calbiochem), and 1% Triton X-100. Samples were subjected to three cycles of sonication of 20 s each before centrifugation and dialysis in TBS, pH 7.4. Expression of the various I domain fragments of the expected molecular weight was confirmed in IPTG-induced bacterial lysates, but not in noninduced samples, by SDS-gel electrophoresis and Coomassie Blue staining. Fibrinogen was purified from fresh frozen human plasma by glycine precipitation (19). The interaction of fibrinogen with ␣ M ␤ 2 on stimulated monocytes and PMN has been described previously (7).
Cell Adhesion-PMN or THP-1 cells were metabolically labeled with 300 Ci of 51 CrNaO 4 (NEN Life Science Products) for 1 h at 37°C, washed, and suspended in serum-free RPMI 1640 at 1 ϫ 10 6 /ml. Twohundred-l aliquots of cell suspension were stimulated with 1 M fMLP (Sigma), treated with the various mAbs at 20 g/ml for 15 min at 22°C, and further incubated with or without fibrinogen (200 g/ml) in 2.5 mM CaCl 2 for 20 min at 22°C. Cells were added to monolayers of HUVEC or ICAM-1 transfectants for 45 min at 22°C, and after washes attached cells were solubilized in 20% SDS with determination of radioactivity in a scintillation ␤-counter. The number of attached cells was calculated by dividing the counts/min observed by the counts/min/cell (21). Alternatively, HUVEC monolayers were preincubated with mAbs IB4, 7E3, or control mAb 14E11 for 30 min at 22°C, washed, and mixed with 51 Cr-labeled, fMLP-stimulated THP-1 cells before determination of fibrinogen-dependent intercellular bridging, as described above. In other experiments, microtiter wells were coated with human or rabbit (Sigma) fibrinogen at 10 g/ml for 18 h at 4°C and blocked with 3% gelatin for 30 min at 37°C. Wells were incubated with fMLP-stimulated 51 Crlabeled THP-1 cells (1 ϫ 10 6 /ml) preincubated with 20 g/ml mAbs 7E3, IB4, or control mAb 14E11 for 10 min at 22°C with determination of cell adhesion after a 45-min incubation at 22°C.
Animal Procedures-Adult New Zealand White rabbits (3-5 kg weight) were given free access to water and rabbit chow and were housed in a facility with alternating light and dark cycles. Animal care complied with the "Principles of Laboratory Animal Care" (National Society for Medical Research) and the "Guide for the Care and Use of Laboratory Animals" (NIH Publication No. 80 -23, revised 1985). The Harvard Medical Area Standing Committee on Animals approved the experimental protocol. For surgical procedures, animals were anesthetized with ketamine (25 mg/kg, Ketalar TM , Parke-Davis) and xylazine (5 mg/kg, Rompun™, Mobay, Shawnee, KS) administered by intramuscular injection, supplemented with intravenous doses of the mixture as needed. The iliofemoral arterial segment was exposed bilaterally via extended groin incisions, and all side branches proximal to the femoral bifurcation were ligated to create an isolated segment (22). The superficial femoral artery was then cannulated, and a 2-French Fogarty balloon embolectomy catheter (American Edwards, Anasco, Puerto Rico) was inserted. The catheter tip was passed into the terminal aorta and the balloon inflated and withdrawn three times to denude the vessel. After removal of the catheter, the superior femoral artery was ligated and antegrade flow re-established via the deep femoral branch. Wounds were closed surgically and the animals allowed to recover. At sacrifice (5-6 days after balloon injury) animals were re-anesthetized as described and systemically heparinized with a 1000-unit intravenous bolus. The groin wounds were reopened and patency of the femoral arteries assessed by direct inspection. The animals were then euthanized with an intravenous overdose of sodium pentobarbital. The abdominal aorta and inferior vena cava were cannulated and the distal arterial tree perfused at 80 -120 mm Hg with 500 ml of heparinized (10 units/ml) lactated Ringer's solution (Baxter Healthcare Corp., Deerfield, IL) with continuous venous drainage. The femoral arteries were re-cannulated with 20-gauge stainless steel catheters and the entire aortofemoral arterial segment excised and placed into sterile PBS.
Ex Vivo Monocyte Adhesion-The freshly excised arterial segments were flushed gently with PBS prewarmed to 37°C, and a microvascular clamp was placed at the origin of each external iliac artery. 51 Cr-Labeled THP-1 cells (10 7 /ml) were stimulated with 1 M fMLP in the presence of 2.5 mM CaCl 2 , equilibrated with control mAb 14E11 or mAb 7E3 (10 -25 g/ml) for 30 min at 22°C, and infused to fill each vessel segment. Paired arteries in each case were treated with mAb 7E3 versus control mAb 14E11. The external surface of the filled arteries was rinsed copiously with PBS and the vessels placed in PBS at 37°C for 1 h. At the end of the incubation, a 5-mm segment of proximal aorta (not exposed to THP-1 cells intraluminally) was excised for measurement of background counts. The clamps were then removed, and each vessel was flushed gently with 3 ml of PBS at 1 ml/min by timed hand injection. The external surface was again rinsed and 5-mm segments cut (two from each iliofemoral segment) for scintillation counting. Segments of the remaining tissue were embedded in optimal cutting temperature compound (Bayer, West Haven, CT) and snap-frozen in 2-methylbutane cooled in liquid nitrogen for immunohistochemical analysis. Tissue was prepared for scintillation counting by carefully mincing each specimen and solubilizing in 1% Triton X-100, 10% SDS overnight at 22°C. The following day an equal volume of scintillant was added, and radioactivity was determined in a scintillation counter.
Immunohistochemistry-Snap-frozen tissue was cut into 6-m-thick cross-sections and adhered to glass slides coated with 0.25% gelatin and 0.025% chromium potassium sulfate (CrK(SO 4 ) 2 , Sigma). Tissue was fixed in acetone at Ϫ20°C for 10 min, quenched in 3% H 2 O 2 , and blocked with 10% normal horse serum. Following each step, slides were washed with TBS containing 135 mM NaCl, 25 mM Tris, 2.6 mM KCl, pH 7.4, plus 1% fetal bovine serum. A mouse anti-fibrinogen antibody characterized in previous studies (19) was applied for 1 h at 22°C, followed by a secondary biotinylated horse anti-mouse antibody (Vector Laboratories Inc. Burlingame, CA) for 2 h at 22°C. Negative controls were prepared in the absence of primary antibody or with nonbinding isotype control (IgG1). Sections were incubated with an avidin-biotin peroxidase complex (ABC kit, Vector) followed by 3-amino-9-carbazole. Slides were counterstained with Mayer's hematoxylin (Sigma) followed by 30% NH 3 OH.
Differential Regulation of Leukocyte Adhesion by mAb 7E3-The effect of mAb 7E3 on ␣ M ␤ 2 recognition of ICAM-1 (1) was first investigated. Preincubation of fMLP-stimulated PMN or monocytic THP-1 with mAb 7E3 or control mAb 14E11 did not reduce the attachment of these cells to monolayers of resting HUVEC or ICAM-1 transfectants (Fig. 3). In contrast, anti-␤ 2 mAb IB4 nearly completely abrogated PMN adhesion to HU-VEC or ICAM-1 transfectants, and significantly inhibited THP-1 cell attachment to either cell type, under the same experimental conditions (Fig. 3).
A potential effect of mAb 7E3 on ␣ M ␤ 2 recognition of fibrinogen (7) was next investigated. In adhesion assays, mAb 7E3 completely inhibited the attachment of fMLP-stimulated THP-1 cells to immobilized human or rabbit fibrinogen (Fig.  4A). Similar results were obtained with anti-␤ 2 mAb IB4, whereas control mAb 14E11 was ineffective (Fig. 4A). The effect of mAb 7E3 on fibrinogen-dependent leukocyte-endothelium bridging (21) was also investigated. Consistent with previous observations (21), fibrinogen enhanced the adhesion of fMLP-stimulated PMN or THP-1 cells to HUVEC or ICAM-1 transfectants by 2-4.4-fold (Fig. 4B, legend), as compared with control incubation reactions in the absence of fibrinogen (Fig.  3). Under these experimental conditions, mAb 7E3 inhibited the fibrinogen-dependent enhancement of PMN or THP-1 cell adhesion to HUVEC or ICAM-1 transfectants (Fig. 4B). In contrast, fibrinogen-independent attachment of PMN or THP-1 cells to either cell type was unaffected by mAb 7E3 (Fig. 4B,   FIG. 2. Epitope mapping of mAb 7E3. A, I domain fragments. The indicated ␣ M I domain fragments were immobilized onto 96-well microtiter plates for 18 h at 4°C, blocked with 3% gelatin, and incubated with 20 g/ml control mAb 14E11, anti-␤ 2 mAb IB4, or mAb 7E3 followed by biotin-conjugated rabbit anti-mouse IgG and streptavidin-alkaline phosphatase plus p-nitrophenyl phosphate. After washes, absorbance was quantitated at A 405 . B, synthetic peptides. The experimental conditions are the same as in A, except that plates were coated with I domain peptides G 127 CPQEDSDIAFLIDGSGSIIPHDF 150 (R1.1), F 150 RRMKEFVSTVMEQLKKSKTLFS 172 (R1.2), control R1.1scrambled GILGFDEGPCIDHASPDISDFQIS (R1.1S), ␤3 integrin peptide V 107 EDYPVDIYYLMDLSYSMKDDL 128 (Beta3), or a control factor X peptide KDGLGEYG (Control) at 5 g/ml in carbonate buffer, pH 9.5. For both panels, data are the mean Ϯ S.E. of three independent experiments. legend). In control experiments, anti-␤ 2 mAb IB4 inhibited fibrinogen-dependent and -independent intercellular adhesion, whereas control mAb 14E11 was ineffective (Fig. 4B). Finally, preincubation of HUVEC monolayers with mAb 7E3 followed by washes and addition of 51 Cr-labeled THP-1 cells failed to reduce fibrinogen-dependent intercellular bridging, thus ruling out a potential role of HUVEC ␣ v ␤ 3 recognition of fibrinogen in this interaction, and in agreement with previous observations (21).
Effect of mAb 7E3 on Monocyte Adhesion to Balloon-injured Arteries-Balloon-mediated injury of external iliac arteries in rabbits resulted in prominent attachment of fMLP-stimulated, 51 Cr-labeled THP-1 cells to the de-endothelialized injured vessel, as compared with noninjured arteries (Fig. 5 and data not shown). Under these experimental conditions, equilibration of THP-1 cells with mAb 7E3 ex vivo nearly completely inhibited the attachment of these cells to balloon-injured vessels, whereas control mAb 14E11 was ineffective (Fig. 5). In four of four animals, mAb 7E3 inhibited monocyte THP-1 cell adhesion to balloon-injured vessels by 53.6 Ϯ 10.6% at 10 g/ml and by 85.9 Ϯ 3.9% at 25 g/ml (n ϭ 2). In immunohistochemical analysis, an anti-fibrinogen antibody strongly reacted with the luminal and medial aspects of the balloon-injured rabbit iliac arteries (Fig. 6), whereas no specific fibrinogen staining was observed in noninjured arteries (not shown). In control experiments, no specific staining of injured vessels was observed in the absence of a primary antibody, under the same experimental conditions (Fig. 6).

DISCUSSION
In this study, we have shown that anti-␣ IIb ␤ 3 mAb 7E3, an integrin antagonist currently used in clinical practice (11), binds a discrete region of ␣ M I domain and inhibits fibrinogenmediated leukocyte adhesion in vitro, and in balloon-injured arteries of rabbits, ex vivo.
Although of paramount importance for normal hemostasis (23), platelet aggregation mechanisms maintained by ␣ IIb ␤ 3 may precipitate thrombus formation and acute ischemic cardiovascular emergencies (5, 6). Among integrin antagonists capable of targeting platelet adherence mechanisms, anti-␣ IIb ␤ 3 mAb 7E3 significantly reduced mortality and emergency procedures in patients undergoing acute coronary intervention (12,13). At the molecular level, it was also shown that mAb 7E3 possessed an unusual pattern of antigen recognition, which included, in addition to ␣ IIb ␤ 3 , the related integrin ␣ v ␤ 3 (9), and the active form of the leukocyte integrin ␣ M ␤ 2 (14,16). This was potentially relevant to the beneficial effect of mAb 7E3 in vivo, because inhibition of ␣ v ␤ 3 reduced neointimal hyperplasia in models of vascular injury (24), and ␣ M ␤ 2 -dependent leukocyte adherence contributed to monocyte recruitment and reperfusion injury (8).
Here, mAb 7E3 bound a cross-reacting epitope in the aminoterminal R1 region Cys 128 -Ser 172 of ␣ M I-domain, which was further narrowed to the R1.1 peptide sequence G 127 C-PQEDSDIAFLIDGSGSIIPHDF 150 . The most salient feature of this motif is the presence of the amino acids Asp 140 -Ser 142 -Ser 144 , which comprise the first group of oxygenated residues of the metal ion-dependent adhesion site (MIDAS) on ␣ M I domain (17). Experimental evidence obtained with homology models, mutagenesis, and divalent ion binding studies suggests FIG. 4. Effect of mAb 7E3 on ␣ M ␤ 2 recognition of fibrinogen. A, adhesion to fibrinogen. Plastic microtiter plates were coated with 10 g/ml human or rabbit fibrinogen for 18 h at 4°C, washed, and blocked with 3% gelatin. 51 Cr-Labeled THP-1 cells (1 ϫ 10 6 /ml) were stimulated with 1 M fMLP, incubated with the indicated mAbs at 20 g/ml for 10 min at 22°C, and added to fibrinogen-coated plates for an additional 45-min incubation at 22°C before determination of cell adhesion. B, fibrinogen-dependent leukocyte-endothelium bridging. The experimental procedures are essentially the same as described in the legend to Fig. 3, except that fMLP-stimulated PMN or THP-1 cells were incubated with the various mAbs and mixed with 200 g/ml fibrinogen, and 2.5 mM CaCl 2 , before addition to monolayers of resting HUVEC or ICAM-1 transfectants. Adherent THP-1 cells to HUVEC or ICAM-1 transfectants in the absence of fibrinogen were 18,937 Ϯ 2,453 and 20,115 Ϯ 6,900, respectively. Adherent PMN to HUVEC or ICAM-1 transfectants in the absence of fibrinogen were 14,183 Ϯ 1531 and 24,233 Ϯ 2201, respectively. For both panels, data are the mean Ϯ S.D. of two independent experiments in duplicate. Cr-labeled THP-1 cells were equilibrated with 25 g/ml control mAb 14E11 or mAb 7E3 for 30 min at 22°C in the presence of 2.5 mM CaCl 2 before injection in the isolated injured vessels. After a 45-min incubation at 22°C, vessels were flushed, minced and radioactivity associated under the various experimental conditions was determined in a scintillation counter. The degree of THP-1 cell adhesion to noninjured aorta in the absence of mAb is indicated (Background). Data are the mean Ϯ S.E. of three independent experiments. that a MIDAS-like motif containing the DXSXS motif, in which X is a nonconserved amino acid, is structurally and functionally present in all integrin ␤ subunits (25)(26)(27)(28). Intriguingly, mAbs raised against this region in ␤ 3 -inhibited fibrinogen binding to the receptor (29), and a synthetic peptide duplicating the MI-DAS-like motif bound ligand and divalent ions (30). Under our experimental conditions, a ␤ 3 -derived peptide containing the MIDAS homology motif failed to associate with mAb 7E3. This is consistent with the inability of mAb 7E3 to recognize the isolated ␤ 3 subunit (31) and suggests that conformational and/or divalent ion-dependent changes in ␣ IIb ␤ 3 are required to form a high affinity antibody binding interface (17). Whether or not mAb 7E3 recognizes MIDAS-like structures in other integrin ␤ subunits is currently not known. However the data presented here suggest that the antigenic accessibility of this shared motif may be modulated by receptor-specific conformational changes and/or the requirement of additional contact site(s) in the ␣/␤ heterodimer. This complex pattern of multiple integrin recognition may not be unique of mAb 7E3, because other function blocking anti-␣ IIb ␤ 3 mAbs, i.e. 25E11, have been shown to cross-react with ␣ M ␤ 2 (32). The next question addressed by this study was the potential physiologic relevance of mAb 7E3 cross-reactivity with ␣ M ␤ 2 . Consistent with the critical role of ␣ M I domain in ligand binding (17,18,33,34), engagement of the mAb 7E3 crossreacting epitope suppressed the receptor recognition of fibrinogen, in agreement with recent observations (16). This resulted in inhibition of leukocyte adherence to immobilized fibrinogen and of fibrinogen-dependent leukocyte-endothelium bridging (21), indistinguishably from bona fide anti-␤ 2 integrin mAb IB4. In contrast, at variance with a recent study (16), mAb 7E3 failed to reduce the ␣ M ␤ 2 recognition of ICAM-1 (1,3). Although differences in protocol may account for this discrepancy, experimental evidence with epitope-mapped mAbs (33), and peptide inhibition studies (35), suggests that the ICAM-1-and fibrinogen-binding sites on ␣ M I domain are physically distinct and nonoverlapping.
An in vivo model of vascular damage further underscored the relevance of mAb 7E3 targeting of monocyte-fibrinogen interaction. In these studies, balloon-mediated injury of the iliofem-oral arteries in rabbits resulted in prominent deposition of fibrinogen as detected by immunohistochemistry and in agreement with previous studies (36). In addition to promoting increased procoagulant activity (37), this translated in our study in prominent monocyte attachment to the injured vessel, in a reaction specifically inhibited by mAb 7E3. Consistent with the importance of leukocyte adherence in vascular disease (38), this suggests that fibrinogen deposited on atherosclerotic lesions (39), and balloon-injured arteries (36), may provide an ideal substrate for leukocyte recruitment. In turn, this may further exacerbate vascular damage by promoting increased IL-1 (40) and tissue factor (41) gene expression, chemotaxis (42), and release of oxidative radicals (43). Whether or not the inhibition of monocyte adhesion by mAb 7E3 contributes to its protective effect in ischemic disease (12, 13) is currently not known. However, the data presented here are consistent with a model in which mAb 7E3 blockade of ␤ 3 integrins on platelets and endothelium, and ␣ M ␤ 2 on leukocytes may simultaneously inhibit multiple cell adherence pathways at the interface between thrombosis and inflammation (44). FIG. 6. Immunohistochemical deposition of fibrinogen on balloon-injured vessels. The experimental conditions are essentially the same as described in the legend to Fig. 5, except that balloon-injured iliac arteries were processed for immunohistochemistry with an antifibrinogen antibody. a, control negative staining in the absence of primary antibody. b and c, immunoreactivity of the anti-fibrinogen antibody with luminal (arrows) and medial aspects of the denuded vessels. Magnification: ϫ 200.