J Biol Chem, Vol. 274, Issue 31, 21609-21616, July 30, 1999

Mechanisms and Consequences of Affinity Modulation of Integrin _V3 Detected with a Novel Patch-engineered Monovalent Ligand^*

Nisar Pampori^§, Takaaki Hato^§, Dwayne G. Stupack^¶, Sallouha Aidoudi, David A. Cheresh^¶, Glen R. Nemerow^¶, and Sanford J. Shattil^**

From the Departments of  Vascular Biology, ^¶ Immunology, and  Molecular and Experimental Medicine, Scripps Research Institute, La Jolla, California 92037

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Integrin _V3 mediates diverse responses in vascular cells, ranging from cell adhesion, migration, and proliferation to uptake of adenoviruses. However, the extent to which _V3 is regulated by changes in receptor conformation (affinity), receptor diffusion/clustering (avidity), or post-receptor events is unknown. Affinity regulation of the related integrin, _IIb3, has been established using a monovalent ligand-mimetic antibody, PAC1 Fab. To determine the role of affinity modulation of _V3, a novel monovalent ligand-mimetic antibody (WOW-1) was created by replacing the heavy chain hypervariable region 3 of PAC1 Fab with a single _V integrin-binding domain from multivalent adenovirus penton base. Both WOW-1 Fab and penton base bound selectively to activated _V3, but not to _IIb3, in receptor and cell binding assays. _V3 affinity varied with the cell type. Unstimulated B-lymphoblastoid cells bound WOW-1 Fab poorly (apparent K_d = 2.4 µM), but acute stimulation with phorbol 12-myristate 13-acetate increased receptor affinity >30-fold (K_d = 80 nM), with no change in receptor number. In contrast, _V3 in melanoma cells was constitutively active, but ligand binding could be suppressed by overexpression of ₃ cytoplasmic tails. Up-regulation of _V3 affinity had functional consequences in that it increased cell adhesion and spreading and promoted adenovirus-mediated gene transfer. These studies establish that _V3 is subject to rapid regulated changes in affinity that influence the biological functions of this integrin.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Integrins mediate cell adhesion and signaling during many developmental, physiological, and pathological processes (1-4). The ₃ integrin family includes _IIb3, often referred to as the fibrinogen receptor, and _V3, the vitronectin receptor. _IIb3 is confined to megakaryocytes and platelets and is required for platelet aggregation through interactions with Arg-Gly-Asp (RGD)-containing adhesive ligands, including fibrinogen and von Willebrand factor (5). _V3 is more widely expressed in proliferating endothelial cells, arterial smooth muscle cells, osteoclasts, platelets, and certain subpopulations of leukocytes and tumor cells (6, 7). The list of cognate ligands for _V3 overlaps that for _IIb3, but includes others, such as osteopontin, matrix metalloproteinase-2, and adenovirus penton base, which do not interact with _IIb3 (6, 8-10). In the adult organism, _V3 has been implicated in processes ranging from wound healing to tumor angiogenesis (11), arterial restenosis (12), osteoporosis (13), tumor progression (14), and adenovirus internalization (8).

One fundamental function of integrins is ligand binding, which in many cases is rapidly regulated by a process variously referred to as "integrin activation," "inside-out signaling," or "affinity/avidity modulation" (15-19). Integrin activation encompasses at least two events: 1) modulation of receptor affinity through conformational changes in the heterodimer and 2) modulation of receptor avidity through facilitation of lateral diffusion and/or clustering of heterodimers (5, 18, 20, 21). The importance of rapid regulated changes in integrin affinity/avidity is easy to appreciate for _IIb3 because platelets must interact productively with fibrinogen or von Willebrand factor only in vascular wounds. Studies of _IIb3 activation have been facilitated by the use of soluble ligands, most notably a multivalent ligand-mimetic antibody called PAC1 and its monovalent Fab fragment, which contain an R(G/Y)D tract in H-CDR3¹ (5, 22). Evidence acquired using these monovalent and multivalent forms of PAC1 indicates that changes in affinity and avidity play complementary roles in _IIb3 function (23).

On the other hand, the significance of inside-out signaling and, in particular, affinity modulation for _V3 has been less certain. The ligand binding function of _V3 has usually been assessed by cell adhesion assays, and these have clearly shown that activation of certain cells leads to _V3-mediated adhesion (24-28). However, adhesion assays can be strongly influenced by post-ligand binding events, including changes in cell shape, that can obscure the precise contributions of affinity or avidity modulation to the overall response. Recently, Byzova and Plow (28, 29) showed that soluble prothrombin, a coagulation protein, can bind directly in an RGD-dependent fashion to _IIb3 in platelets and _V3 in cultured human endothelial cells. Notably, prothrombin binding to _V3 could be increased by MnCl₂, a general activator of integrins, and by phorbol 12-myristate 13-acetate, an activator of protein kinase C. Taken together, these results indicate that _V3 has the potential to be regulated at the level of ligand binding. However, the specific mechanisms, relative contributions, and biological consequences of affinity and avidity modulation of _V3 remain to be established. The distinction between affinity and avidity regulation is not academic because the underlying mechanisms and effects on cell function are likely to be different (5, 18, 19). Furthermore, mechanistic insights into the regulation of ligand binding could facilitate current efforts aimed at developing drugs that inhibit or stimulate _V3 function in vivo.

The purpose of these studies was to determine whether _V3 is subject to affinity modulation and, if so, to explore the potential pathophysiological implications of such regulation. To accomplish this task, we characterized the binding of soluble monovalent and multivalent ligands to _V3 in several cell types, reasoning that a monovalent ligand would be sensitive to affinity modulation and a multivalent ligand would be sensitive to both affinity and avidity modulation (18, 23, 30). Penton base, a coat protein from adenovirus type 2, was selected as a multivalent ligand because each of its five subunits contains a 50-amino acid RGD tract that mediates virus internalization through _V integrins (8). A novel monovalent ligand called WOW-1 Fab was created by replacing the H-CDR3 of PAC1 Fab with a single integrin-binding domain of penton base. This switched the selectivity of the Fab fragment from activated _IIb3 to activated _V3, enabling a direct assessment of _V3 affinity state. These studies establish that _V3 is subject to affinity regulation, with direct implications for the anchorage-dependent functions of _V3 and for adenovirus-mediated gene delivery.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Preparation of Soluble _V3 Ligands-- Recombinant penton base from adenovirus type 2 was baculovirus-expressed in Trichoplusia Tn 5B1-4 insect cells and purified as described previously (8). The purified protein migrated as a single ~325-kDa band on native polyacrylamide gels and an ~80-kDa band on SDS-polyacrylamide gels. Penton base was conjugated to Alexa-488 to form Alexa-penton base (aPB)¹ according to the manufacturer's instructions (Molecular Probes, Inc., Eugene, OR). Purified human fibrinogen was obtained from Enzyme Research Laboratories (South Bend, IN) and labeled with fluorescein isothiocyanate (31).

WOW-1 Fab was created by replacing the 19-amino acid H-CDR3 of antibody PAC1 Fab (22) with the 50-amino acid _V integrin-binding domain from adenovirus type 2 penton base (32) by splice-overlap polymerase chain reaction using oligonucleotides PB-For (5'-ACACAGCCATATATTACTGTGCCAGAGCGGAAGAGAACTCCAACGCG), PB-Rev (5'-ACTGAGGTTCCTTGACCCCACGCAGCGGGGGCGGCAGCTTCTGC), Pac1-For (5'-GCGCGGGAGATCTCAGGTGCAGCTGAAGCAGTCAGGA), and Pac1-Rev (5'-GGCGCATGACCGGTACAATCCCTGGGCACAATTTTCTTG). The resulting WOW-1 fd DNA fragment was digested with BglII/AgeI and cloned into a Drosophila expression vector, pMT/BiP/V5-HisB (Invitrogen, Carlsbad, CA), which contains the Drosophila metallothionein promoter and BiP secretion signal and places a His₆ tag at the C terminus of fd DNA. Similarly, PAC1 containing NcoI and AgeI sites was amplified by polymerase chain reaction with -For (5'-GGCGCGGGAGATCTCCATGGGATGTTTTGATGACCCAAACTCCA) and -Rev (5'-GGCGCATGACCGGTACACTCATTCCTGTTGAAGCTCTTG) and cloned into pMT/BiP/V5-HisB. Nineteen µg of WOW-1 fd and PAC1 DNA in pMT/BiP/V5-HisB were cotransfected with 1 µg of selection vector (pCoHYGRO, Invitrogen) into Drosophila melanogaster S2 cells by calcium phosphate precipitation. Stable S2 cell lines were selected with hygromycin B and screened for secretion of WOW-1 Fab after a 36-72-h induction with 500 µM CuSO₄.

WOW-1 Fab was purified from 250-1000 ml of serum-free medium by column chromatography on Ni²⁺-nitrilotriacetic acid (QIAGEN Inc., Chatsworth, CA). Typical yields were 2-5 mg/liter with a purity of 90% as estimated on SDS gels stained with silver or Coomassie Blue. WOW-1 Fab migrated as a single ~58-kDa band on nonreduced SDS gels and reacted on Western blots with a monoclonal antibody specific for a linear epitope in the integrin-binding domain of penton base (33) and with affinity-purified goat anti-mouse -chain (BIOSOURCE International, Camarillo, CA). After reduction, WOW-1 Fab migrated as an ~33-kDa fd chain and an ~25-kDa -chain. There was no evidence of fd or -homodimers. As with PAC1 Fab (22), the relative migration of WOW-1 Fab on a Sephadex G-200 column indicated that it was monomeric and therefore monovalent in aqueous solution.

Mammalian Cells and DNA Transfections-- cDNAs encoding full-length human _V and ₃ were subcloned into pcDNA3 and pCDM8, respectively, and 2 µg of each were transfected into CHO-K1 cells to obtain transient and stable transfectants as described (34). Stable transfectants surviving antibiotic selection were further screened for high _V3 expression by single-cell fluorescence-activated cell sorting using the _V3-specific monoclonal antibody LM609 (35). CHO cells stably expressing wild-type human _IIb3 and _V3(D723R) were described previously (34, 36). M21-L is a clone of the human melanoma cell line M21 that lacks the _V subunit (37). _V3-M21-L cells were produced by transient transfection of M21-L cells with 2 µg each of _V/pcDNA3 and ₃/pCDM8 using Superfect (QIAGEN Inc.). CS-1 is a hamster melanoma cell line that does not express _V3 or _V5 because it does not synthesize the ₃ or ₅ subunits. _V3-CS-1 cells stably expressing hamster _V and human ₃ were obtained by transfection of CS-1 cells with human ₃ (38). JY is an immortalized human B-lymphoblastoid cell line that expresses _V3, but not _V5 (24, 39).

Analysis of Cell-surface Integrin Expression-- Cells were suspended in incubation buffer (137 mM NaCl, 2.7 mM KCl, 3.3 mM NaH₂PO₄, 3.8 mM HEPES, 1 mM MgCl₂, 5.5 mM glucose, and 1 mg/ml bovine serum albumin, pH 7.4) and incubated for 30 min on ice with a monoclonal antibody (10 µg/ml) specific for _V3 (LM609), _IIb3 (D57) (34), or _V5 (P1F6) (40). After washing, the cells were incubated for another 30 min on ice with fluorescein isothiocyanate-conjugated goat anti-mouse IgG (H + L chain-specific; BIOSOURCE International), washed again, and analyzed on a FACSCalibur flow cytometer (Becton Dickinson, Mountain View, CA) (23). As a negative control, samples were incubated with the secondary antibody alone.

Ligand Binding Assays-- Binding of aPB, WOW-1 Fab, and fluorescein isothiocyanate-fibrinogen to cells was assessed by flow cytometry. Typically, cells were cultured overnight in low serum medium (e.g. 0.5% fetal bovine serum) and resuspended in incubation buffer at 1-1.5 × 10⁷ cells/ml, and 4-6 × 10⁵ cells were incubated with one of these ligands for 30 min at room temperature in a final volume of 50 µl. As indicated, some samples were also incubated in the presence of one or more of the following reagents: antibody AP5 ascites (1:50) to activate ₃ integrins (41); 0.25 mM MnCl₂ to activate integrins (18); 2 mM RGDS or 5 mM EDTA to specifically block ligand binding to integrins; 50 µM cRGDfV, a selective _V integrin antagonist (Peninsula Laboratories, Inc., Belmont, CA); 5 µM Integrilin, a selective _IIb3 antagonist (42); or a 100 µg/ml concentration of the function-blocking antibody LM609 or P1F6. In some experiments, ligand binding and _V3 expression were measured simultaneously by incubation of cells with ligands in the presence of biotin-SSA6 (7 µg/ml), a non-function-blocking anti-₃ monoclonal antibody (22). After 30 min at room temperature, cells were washed and incubated with phycoerythrin-streptavidin (1:25 final dilution; Molecular Probes, Inc.) for 20 min on ice. In the case of WOW-1 Fab, an Alexa-conjugated anti-His₆ monoclonal antibody (Accurate Chemical and Scientific Corp., Westbury, NY) was added at this stage (50 µg/ml). Cells were washed and resuspended in 0.5 ml of incubation buffer containing 2 µg/ml propidium iodide (Sigma). Ligand binding (FL1 channel) was analyzed immediately on the gated subset of live cells (propidium iodide-negative, FL3) that were strongly positive for _V3 expression (FL2). Binding isotherms were subjected to nonlinear least-squares regression analysis using an equation for one-site binding (Prism 2.0 software, GraphPAD Software for Science, San Diego, CA). Two-tailed p values for paired samples were obtained by Student's t test.

To examine the effects of overexpression of isolated integrin cytoplasmic tails on ligand binding to _V3, _V3-CS-1 cells were transfected with a mammalian expression plasmid encoding Tac-₁, Tac-₃, or Tac-₅, using Fugene-6 transfection reagent (Roche Molecular Biochemicals) (43, 44). Forty-eight h after transfection, cells were suspended in incubation buffer at 1.5 × 10⁶/ml and incubated for 30 min at room temperature with 150 nM aPB or 425 nM WOW-1 Fab in the presence or absence of 5 mM EDTA. After washing, cells were incubated for an additional 30 min on ice with 2.5 µg/ml biotinylated anti-Tac monoclonal antibody (7G7B6), followed by incubation with phycoerythrin-conjugated anti-mouse IgG and (when WOW-1 Fab was present) 50 µg/ml Alexa-conjugated anti-His₆ monoclonal antibody. Ligand binding was analyzed on the gated subset of live cells strongly positive for Tac expression. In parallel tubes, cells were co-stained with SSA6 and anti-Tac antibody to quantitate _V3 expression in the Tac-positive cells.

Binding of WOW-1 Fab to purified _V3 receptors from human placenta and _IIb3 from human platelets was measured by enzyme-linked immunosorbent assay in the presence of 50 µM CaCl₂, MgCl₂, and MnCl₂. Nonspecific binding was determined in the presence of 2 mM RGDS (22).

Cell Adhesion Assays-- Immulon-2 microtiter wells (Dynex Laboratories, Chantilly, VA) were coated with unlabeled penton base (1-100 ng/well) overnight at 4 °C, followed by blocking with 20 mg/ml bovine serum albumin. CHO cells stably expressing _V3 were labeled with 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein/acetoxymethyl ester (Molecular Probes, Inc.), and cell adhesion to immobilized penton base was quantitated by cytofluorometry at 485/530 nm (23).

Adenovirus-mediated Gene Delivery-- CS-1 and _V3-CS-1 cells (10⁵ cells) were suspended for 5 min at room temperature in 100 µl of incubation buffer. In some cases, 2.5 mM MnCl₂ was also present to induce maximal integrin activation. Then, replication-deficient adenovirus type 5 encoding green fluorescent protein (GFP) was added to the cell suspension at a multiplicity of infection (m.o.i.) of 50 or 500 (45). After 1 h at 37 °C, virus not internalized was digested by incubation of the cells with 0.03% trypsin and 0.35 mM EDTA for 5 min at 37 °C. After 72 h, GFP expression was quantitated by flow cytometry.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Interaction of a Novel Monovalent Ligand with Integrin _V3-- To document and study the significance of affinity modulation of _V3, we set out to develop a monovalent reporter ligand analogous to the activation-dependent anti-_IIb3 antibody PAC1 Fab. We reasoned that swapping the 50-amino acid RGD motif from adenovirus penton base with the H-CDR3 of PAC1 Fab might convert the antibody into a suitable _V3 ligand. To test this idea, preliminary binding studies were conducted with the new antibody, designated WOW-1 Fab, using purified integrins in the presence of 50 µM MnCl₂, which activates integrins by a direct effect on the extracellular domain (18). WOW-1 Fab bound to purified _V3 and, to a lesser extent, to purified _V5. Binding was half-maximal at 40 nM Fab fragment and was blocked by >95% by 2 mM RGDS or 5 mM EDTA. In contrast, there was no detectable binding of WOW-1 Fab to purified _IIb3 at antibody concentrations as high as 2 µM, even though the parent antibody, PAC1 Fab, bound half-maximally to _IIb3 at 50 nM. These results indicate that the reengineering of PAC1 Fab converted it from an activation-dependent _IIb3 antibody into an antibody that reacts with activated _V3.

To determine if WOW-1 Fab reacted preferentially with activated _V3 in cells, Fab fragment binding was compared with that of multivalent penton base using CHO cells stably transfected with human _V3 (_V3-CHO cells). Flow cytometric analysis showed that the surface of these cells expressed large amounts of _V3, modest amounts of _V5, and no detectable _IIb3 (Fig. 1A). When aPB or WOW-1 Fab was incubated with the cells over a range of ligand concentrations (5-1000 nM) and for various periods of time at room temperature, specific ligand binding, defined as that inhibitable by 2 mM RGDS or 5 mM EDTA, reached a steady state by 30 min, and nonspecific binding accounted for 15% of total binding. Therefore, all subsequent binding studies were carried out under these conditions. aPB and WOW-1 Fab bound specifically but at low levels to unstimulated _V3-CHO cells. However, direct activation of _V3 by anti-₃ antibody AP5 caused a significant increase in the binding of both ligands (p < 0.01) (Fig. 1B).

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Binding of soluble Alexa-penton base and WOW-1 Fab to CHO cells expressing V3. In A, V3-CHO cells or parental CHO cells were incubated with primary antibodies specific for V3 (LM609), IIb3 (D57), or V5 (P1F6), and antibody binding was detected with fluorescein isothiocyanate-labeled secondary antibody as described under "Experimental Procedures." Cells stained with secondary antibody (2°Ab) only were used as a negative control. For comparison, antibody binding to parental CHO cells was also studied. In B, the V3-CHO cells were incubated with either 75 nM aPB or 106 nM WOW-1 Fab for 30 min at room temperature in the absence or presence of a 1:50 dilution of antibody AP5 ascites to activate V3 or 5 mM EDTA to inhibit specific ligand binding. Then, binding of aPB and WOW-1 Fab was measured by flow cytometry as described under "Experimental Procedures." The data represent specific ligand binding, defined as that inhibited by EDTA, and are presented as the means ± S.E. of three independent experiments. Similar results were obtained if V3 was stimulated with the purified Fab fragment of another activating antibody (LIBS6) instead of antibody AP5 ascites. Asterisks indicate that ligand binding was significantly greater in the presence than in the absence of antibody AP5 (p < 0.01).

To assess the selectivity of these ligands for _V3 in this system, the effect of various function-blocking compounds was studied. Binding of aPB and WOW-1 Fab in the presence of antibody AP5 was inhibited 85% by 2 mM RGDS or 50 µM cRGDfV, a cyclic peptide selective for _V integrins (Fig. 2) (46). On the other hand, a cyclic peptide selective for _IIb3 (Integrilin) inhibited ligand binding by <20%, even at a concentration (1 µM) 100-fold higher than necessary to prevent fibrinogen or PAC1 binding to platelet _IIb3 (42). Furthermore, the _V3 function-blocking antibody LM609 (100 µg/ml) inhibited ligand binding by >70%, whereas the _V5 function-blocking antibody P1F6 had no such effect (data not shown). In addition, neither aPB nor WOW-1 Fab bound detectably to resting or thrombin-stimulated human platelets, which express >50,000 _IIb3 receptors, but <500 _V3 receptors/cell (7). Collectively, these results indicate that a monovalent ligand, WOW-1 Fab, and a multivalent ligand, aPB, are sensitive to the activation state of _V3 and that they do not recognize _IIb3. Thus, WOW-1 Fab may be a suitable reporter for changes in _V3 affinity. Since WOW-1 Fab (and aPB) also recognizes _V5, particular efforts were made in the experiments that follow to utilize cells that express _V3, but little or no _V5.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effect of integrin inhibitors on binding of aPB and WOW-1 Fab to V3-CHO cells. Ligand binding was carried out as described in the legend to Fig. 1 in the presence of antibody AP5 ascites (1:50) and an integrin inhibitor, as indicated. EDTA was present at 5 mM, RGDS at 2 mM, cRGDfV at 50 µM, and Integrilin at 1 µM. Data are plotted as a percentage of the value for the antibody AP5-treated sample in the absence of an inhibitor and represent the means ± S.E. of three experiments.

Affinity of _V3 Can Be Regulated by Inside-out Signals-- To determine if _V3 is susceptible to affinity modulation by inside-out signals, the binding of WOW-1 Fab to JY B-lymphoblasts was studied. These cells were selected because they express _V3, but not _V5, and they adhere rapidly to vitronectin in response to activation of protein kinase C by phorbol 12-myristate 13-acetate (24, 39). Incubation of JY cells for 15 min with 100 nM phorbol 12-myristate 13-acetate caused a significant increase in specific binding of aPB (2.7 ± 0.2-fold increase; p < 0.05), consistent with an increase in _V3 affinity and/or avidity. Furthermore, phorbol 12-myristate 13-acetate caused a 2.4 ± 0.1-fold increase in the binding of WOW-1 Fab (p < 0.05). Neither response was increased further by activating antibody AP5 (Fig. 3A). Phorbol 12-myristate 13-acetate did not increase the surface expression of _V3, as measured by antibody LM609. To determine whether the changes in WOW-1 Fab binding reflected changes in _V3 affinity, ligand binding was analyzed over a range of antibody concentrations. Unstimulated JY cells exhibited a very low affinity for WOW-1 Fab (apparent K_d = 2600 ± 700 nM; mean ± S.E.) and a value for maximal binding of 24.8 ± 4.1 arbitrary fluorescence units (Fig. 3B). In marked contrast, JY cells stimulated with phorbol 12-myristate 13-acetate exhibited a >30-fold increase in binding affinity (apparent K_d = 80 ± 18 nM) with no change in maximal binding (23.5 ± 1.1 units). This effect was prevented if the cells were first depleted of metabolic energy by a 30-min preincubation with 0.2% sodium azide and 4 mg/ml 2-deoxy-D-glucose. These results establish that energy-dependent inside-out signals can regulate the ligand binding affinity of _V3.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   V3 is susceptible to affinity modulation by inside-out signals. In A, JY lymphoblastoid cells were incubated in the presence of either 75 nM aPB or 425 nM WOW-1 Fab for 15 min without an agonist (No Tx), with 100 nM phorbol 12-myristate 13-acetate (PMA), or with phorbol 12-myristate 13-acetate plus antibody AP5 ascites (1:50). Then, specific ligand binding was determined by flow cytometry. Data are the means ± S.E. of three experiments. Asterisks denote a significant difference compared with the No Tx sample (p < 0.05). In B, binding of WOW-1 Fab to JY cells was examined over a range of Fab fragment concentrations. The data are plotted as specific (RGDS-inhibitable) binding and were subjected to nonlinear regression analysis for binding to a single site. Values for apparent Kd and maximal binding are presented under "Results." The curves are computer-generated best fits of the data. Goodness of fit (R2) values ranged from 0.94 to 1.00.

Determinants of _V3 Activation State-- Additional experiments were performed to identify factors that influence _V3 affinity using readily transfectable cell lines that stably express human _V3. _V3 on vascular cells may encounter multiple ligands simultaneously during the process of wound healing. Therefore, we wondered if the affinity/avidity of _V3 differed for various ligands. Equilibrium binding of aPB, WOW-1 Fab, and the adhesive ligand fibrinogen was compared in _V3-CHO cells. As summarized in Table I, each ligand bound specifically to approximately the same total number of receptors in unstimulated _V3-CHO cells. However, the affinity/avidity of _V3 for fibrinogen was ~15-fold lower than that for aPB, despite the fact that both ligands are multivalent and similar in molecular mass. Activation of _V3 with antibody AP5 increased the binding affinity/avidity for both ligands, but it had no effect on maximal binding (Table I). On the other hand, despite the differences in valency between aPB and WOW-1 Fab, their binding constants were similar. Overall, these results show that _V3 can interact differentially with macromolecular ligands and that the affinity state of the receptor is one determinant of such interactions.

View this table: [in this window] [in a new window]   Table I Binding of different ligands to V3 expressed in CHO cells Ligand binding was determined by flow cytometry, and binding isotherms were analyzed as described under "Experimental Procedures" and in the legend to Fig. 3. Data represent the combined results of three independent experiments with each ligand. Maximum binding (Bmax) is expressed in arbitrary fluorescence units. Goodness of fit (R2) values ranged from 0.93 to 1.00.

In circulating platelets, the "basal" activation state of _IIb3 must remain low to prevent thrombosis. However, this requirement may not pertain to all cells that express _V3. Therefore, ligand binding was studied simultaneously in _V3-CHO cells and in two unrelated melanoma cell lines, _V3-M21-L and _V3-CS-1, to assess cell type-specific variations in the basal activation state of _V3. To control for minor variations in _V3 expression between the cell lines, ligand binding was expressed on a "per receptor" basis using anti-₃ antibody SSA6 to quantitate receptor expression. Unstimulated _V3-M21-L cells bound significantly more aPB than did _V3-CHO cells (p < 0.01). This difference was maintained even after further activation of _V3 with antibody AP5 (p < 0.05) (Fig. 4). Similar results were obtained with _V3-CS-1 cells instead of _V3-M21-L cells and with WOW-1 Fab instead of aPB (data not shown). Taken together with the marked differences observed in the binding of WOW-1 Fab to unstimulated JY lymphoblasts and _V3-CHO cells (Fig. 3B and Table I), these results indicate that the basal activation state of _V3 varies with the cell type.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Comparison of aPB binding to V3-CHO cells and V3-M21-L melanoma cells. Binding of aPB (75 nM) to each cell line was carried out as described in the legend to Fig. 1. Specific aPB binding is expressed on a per receptor basis as the mean fluorescence intensity (mfi) of aPB binding divided by the mean fluorescence intensity of SSA6 binding. Each bar represents the mean ± S.E. of four experiments. Single and double asterisks denote p values of <0.01 and <0.05, respectively, for the difference between the CHO and melanoma cells.

Integrin cytoplasmic tails have been implicated in affinity/avidity modulation of several integrins (47), but there is no direct information about their role in regulating ligand binding to _V3. Certain point mutations or truncations of the ₃ cytoplasmic tail, such as ₃(D723R), result in constitutive activation of _IIb3 in CHO cells (34, 36). To determine whether _V3 is affected by such a modification, ligand binding to _V3(D723R) was assessed. This mutant was stably expressed in CHO cells to approximately the same level as wild-type _V3 (Fig. 5A). However, unstimulated _V3(D723R)-CHO cells bound significantly more aPB than unstimulated _V3-CHO cells (p < 0.01), equivalent to the amount of aPB bound to _V3-CHO cells treated with antibody AP5 (Fig. 5B). A second _V3(D723R) clone gave the same results, and similar results were obtained using WOW-1 Fab instead of aPB. Thus, a structural change in the ₃ cytoplasmic tail can be propagated to the extracellular domains of _V3 to influence ligand binding affinity.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Effect of an activating mutation in the 3 integrin cytoplasmic tail on the binding of penton base to V3. In A, stable CHO cell lines expressing either V3 or V3(D723R) were stained with anti-3 antibody SSA6 and phycoerythrin-streptavidin to assess surface expression of V3. In B, specific binding of aPB (75 nM) was studied as described in the legend to Fig. 1. aPB binding is expressed on a per receptor basis. Data represent the means ± S.E. of four experiments. The asterisk denotes a difference between V3 and V3(D723R) at the p < 0.01 level. For comparison, the corresponding value for aPB binding to antibody AP5-treated V3-CHO cells was 0.034 ± 0.002. 1°Ab, primary antibody; mfi, mean fluorescence intensity.

The activation state of certain integrins, such as _IIb3 and ₅1, can be suppressed in a dominant-inhibitory fashion by overexpression of isolated ₃ or ₁ cytoplasmic tails, but not by ₅ tails (43, 44, 48). To determine if _V3 is subject to this type of suppression, _V3-CS-1 cells were transiently transfected with chimeric constructs consisting of the ₃, ₁, or ₅ cytoplasmic tails fused at their N termini to the extracellular and transmembrane domains of the Tac subunit of the interleukin-2 receptor, which was used to target the tails to the vicinity of the plasma membrane. Despite similar levels of expression of the chimeras, Tac-₃ and Tac-₁ caused a significant reduction in specific binding of aPB and WOW-1 Fab when compared with Tac-₅ (p < 0.01) (Fig. 6, A and B). In contrast, none of these tail chimeras affected surface expression of _V3 (Fig. 6C). Since the isolated  tails may bind proteins that normally interact with integrins (43), these results suggest that _V3 may be regulated by direct interactions with intracellular proteins.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Effect of overexpression of isolated integrin cytoplasmic tails on ligand binding to CS-1 melanoma cells expressing V3. As described under "Experimental Procedures," V3-CS-1 cells were transiently transfected with the Tac-5, Tac-1, or Tac-3 chimera. Forty-eight h after transfection, the cells were incubated for 30 min at room temperature with 150 nM aPB (A) or 425 nM WOW-1 Fab (B) in the presence or absence of 5 mM EDTA. The cells were stained with anti-Tac antibody and phycoerythrin-conjugated anti-mouse IgG in order to set a live gate on the Tac-expressing cells, and specific binding of aPB and WOW-1 Fab was measured by flow cytometry. C shows that the Tac constructs had no effect on expression levels of V3 as monitored with anti-3 antibody SSA6. Data represent the means ± S.E. of three experiments. The asterisks indicate that ligand binding in the presence of Tac-1 or Tac-3 was significantly less than with Tac-5 (p < 0.01).

Functional Consequences of Affinity Modulation of _V3-- To determine whether changes in receptor affinity affect the adhesive function of _V3, the adhesion of _V3-CHO cells to immobilized penton base was quantitated. Adhesion was dependent on the coating concentration of penton base and was half-maximal at 30-40 ng/well (Fig. 7). Activation of _V3 by antibody AP5 led to a 7-fold leftward shift in the dose-response curve such that half-maximal adhesion now occurred at ~5 ng of penton base/well. Treatment of the cells with 1 mM MnCl₂ caused an even further shift in the dose-response curve, either because it induced a more profound effect on _V3 or it activated additional _V integrins (Fig. 7). Analysis of adherent cells by light microscopy showed that they had become fully spread by 90 min. Thus, affinity modulation of _V3 promotes both cell adhesion and post-ligand binding responses, such as cell spreading.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Effect of V3 activation on the adhesion of V3-CHO cells to penton base. As described under "Experimental Procedures," microtiter wells were coated with penton base, and the adhesion of V3-CHO cells was studied for 90 min at 37 °C either with no additive () or with antibody AP5 ascites (1:50; ) or MnCl2 (0.25 mM; ). Some aliquots were also incubated with 50 µM cRGDfV under each of these conditions (, ×, and *, respectively) to assess whether cell adhesion was dependent on the presence of V integrins. This experiment is representative of three so performed.

Adenoviruses utilize _V integrins to enter cells and are a common gene delivery vector. Therefore, we tested whether changes in _V3 affinity could influence adenovirus-mediated gene transfer. Recombinant adenovirus containing cDNA encoding GFP was incubated with CS-1 melanoma cells at m.o.i. of 50 and 500, and subsequent cellular expression of GFP was taken as a marker for infection and gene transfer. CS-1 cells were chosen because they do not express _V5, thus potentially restricting adenovirus internalization through stably expressed _V3. When parental cells without _V3 were incubated with virus for 60 min and monitored for infection 72 h later, they exhibited a relatively low level of GFP expression. Unstimulated _V3-CS-1 cells exhibited a higher level of GFP expression, particularly at the higher m.o.i. (Fig. 8A). However, if incubation of _V3-CS-1 cells with virus was carried out in the presence of 2.5 mM MnCl₂ to activate _V3, the cells subsequently exhibited a much greater increase in GFP expression at the lower m.o.i. (p < 0.01) (Fig. 8, A and B, black bars). MnCl₂ had no effect on GFP expression in the parental CS-1 cells. Enhanced GFP expression in cells containing activated _V3 was blocked if the cells were preincubated with an excess of WOW-1 Fab (1.7 µM) before the addition of virus (Fig. 8B, gray bar). Thus, adenovirus-mediated gene transfer is directly affected by affinity modulation of _V3.

View larger version (22K): [in this window] [in a new window]   Fig. 8.   Effect of V3 expression and activation on adenovirus-mediated gene delivery. In A, parental CS-1 cells (No V3) and V3-CS-1 cells were incubated for 1 h with an adenovirus vector encoding GFP at a multiplicity of infection of 50 or 500. In addition, aliquots of the V3-CS-1 cells were incubated with virus in the presence of 2.5 mM MnCl2 to induce maximal integrin activation. Viral infection and gene delivery were assessed 72 h later by quantitating cellular expression of GFP by flow cytometry. A depicts a single experiment, and B shows the means ± S.E. of three experiments conducted at a m.o.i. of 50. The gray bar shows the effect of preincubating V3-CS-1 cells with 1.7 µM WOW-1 Fab for 20 min before addition of virus. mfi, mean fluorescence intensity.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The purpose of this study was to determine if the affinity of _V3 for ligands is subject to acute modulation in cells. If this were the case, it could have significant clinical implications since numerous physiological and pathological events are mediated by _V3. We reasoned that the most direct way to evaluate _V3 affinity was to employ a soluble, monovalent macromolecular ligand. Since no such ligand exists, we engineered one. First, we determined that a soluble multivalent ligand, adenovirus penton base, binds to _V3 in an activation-dependent manner. Then, in an example of patch engineering (49), we grafted a small portion of penton base, the 50-amino acid _V integrin-binding domain, into the H-CDR3 of PAC1, a monovalent Fab antibody that is specific for activated _IIb3. The resulting monovalent Fab antibody, WOW-1, retained the activation-dependent characteristics of PAC1 and penton base and interacted with _V3, but not _IIb3. Using WOW-1 Fab to study _V3, the following major conclusions were reached. 1) The basal affinity state of _V3 varies among cell types, being extremely low in lymphoid cells and higher in melanoma cell lines. 2) _V3 is subject to rapid affinity modulation by inside-out signals, including those downstream of protein kinase C. 3) At least some of the cellular signals that regulate _V3 affinity appear to converge at the cytoplasmic tails of the integrin. 4) Affinity modulation has direct functional consequences, both for the adhesion and signaling functions of _V3 and for adenovirus-mediated gene transfer.

_V3 Is Subject to Affinity Modulation-- Studies with numerous integrins have suggested that no single signaling pathway or mechanism is likely to regulate ligand binding to all integrins (5, 15-18, 50). This means that mechanisms of integrin activation relevant to _IIb3 do not necessarily apply to _V3. Unfortunately, in the case of _V3, it has been difficult to directly assess the precise contributions and mechanisms of affinity and avidity modulation because most preferred ligands, including vitronectin and osteopontin, are predominantly matrix-associated or difficult to work with in soluble form. Moreover, whereas cellular agonists can stimulate rapid _V3-dependent adhesion and/or migration of platelets, leukocyte cell lines, endothelial cells, and vascular smooth muscle cells (24-28), the adhesion end point does not reveal the relative contributions of ligand binding and post-ligand binding events to the overall response.

Recently, Byzova and Plow (29) and Byzova et al. (51) showed that soluble prothrombin can interact with both _IIb3 and _V3 in a cell activation-dependent manner. In particular, the binding of prothrombin to _V3 in human umbilical vein endothelial cells was increased markedly by treatment with MnCl₂ or a phorbol ester. The results with phorbol ester established that signals triggered by protein kinase C can increase the ligand binding function of _V3 in these cells. They also raised new questions that have now been addressed in the present work. Does acute regulation of _V3 affect the binding of soluble ligands other than prothrombin, and is it relevant to a wider variety of cell types? Does regulation occur at the level of _V3 affinity, or do cells primarily control avidity? If _V3 affinity is regulated, what cellular functions are affected by this process?

Adenovirus penton base is a homopentamer, and cryoelectron micrographs indicate that each of its five RGD tracts is situated at the apex of one of five 22-Å protrusions (33). Thus, penton base may be able to interact simultaneously with two or more _V3 receptors. Consequently, this ligand should be sensitive to both affinity modulation (e.g. conformational change) and avidity modulation (e.g. lateral diffusion/clustering) of _V3. In contrast and as shown experimentally for PAC1 Fab and _IIb3, a monovalent ligand is likely to be more sensitive to affinity than avidity modulation (23, 30). We found that both penton base and WOW-1 Fab bound specifically and saturably to _V3 in an activation-dependent manner. This was true whether _V3 was stimulated from outside the cell by an activating antibody or MnCl₂ or from inside the cell by biochemical signals or by virtue of an activating mutation in the ₃ cytoplasmic tail. Thus, the results with WOW-1 Fab establish that _V3 is susceptible to affinity modulation. It is important to emphasize, however, that these results do not exclude a significant role for lateral diffusion and clustering of _V3 receptors. Indeed, several other integrins have been shown to undergo diffusion and clustering within the plane of the plasma membrane (17, 20, 21, 52). Even in the case of _IIb3, the prototypic integrin regulated by affinity changes, receptor clustering facilitates irreversible ligand binding and is required for outside-in signaling (23).

Like penton base and WOW-1 Fab, soluble fibrinogen also bound to _V3-CHO cells in an activation-dependent manner (Table I). On the other hand, Byzova and Plow (28) showed that adhesion of endothelial cells to immobilized fibrinogen via _V3 did not require cell activation. This suggests that soluble fibrinogen is an activation-dependent ligand for _V3, whereas solid-phase fibrinogen is an activation-independent ligand. A similar situation pertains to the interaction of fibrinogen with _IIb3 in platelets (53). Thus, interactions between ₃ integrins and fibrinogen are governed by the physical state and conformation of both the receptor and the ligand. Of note, the apparent K_d for soluble fibrinogen binding to unstimulated _V3-CHO cells (9200 nM) was much different from the apparent K_d for penton base or WOW-1 Fab (~500 nM), but this difference narrowed after receptor activation (Table I). Since _V3 in vascular cells likely encounters multiple cognate ligands during wound healing, these results indicate that the relevant biological interaction will depend on several factors, including the concentration and physical state of the ligands and the activation state and membrane density of _V3.

Regulation of Affinity Changes in _V3-- Binding of penton base and WOW-1 Fab to