Introduction

Top Abstract Introduction Procedures Results Discussion References

The platelet integrin, _IIb3, (also termed fibrinogen receptor and glycoprotein IIb/IIIa) plays a pivotal role in the final pathway for platelet aggregation and thrombus formation (Marguerie et al., 1980). In response to platelet activation, _IIb3 binds fibrinogen, which in turn cross-links other platelets, forming platelet rich thrombi (Frojmovic et al., 1991). Antagonists of _IIb3 may be useful for the treatment and prevention of the acute coronary syndromes of myocardial infarction and unstable angina as well as the acute-phase response to percutaneous coronary interventions (Topol, 1995).

_IIb3 interacts with several adhesive proteins in addition to fibrinogen, including von Willebrand factor, vitronectin and fibronectin (Plow and Ginsberg, 1988). The recognition site common to the adhesive proteins that bind to _IIb3 contains the tripeptide sequence RGD (Pytela et al., 1996). In addition, fibrinogen has a second receptor recognition site found at the C-terminus of the fibrinogen -chain (HHLGGAKQAGDV) (Kloczewiak et al., 1982). Recombinant fibrinogen with a 20-amino acid insert in place of the terminal AGDV sequence was unable to support platelet aggregation (Farrell et al., 1992).

Small peptides containing the sequence RGD have been shown to be _IIb3 antagonists. Increase in potency was observed when the RGD sequence was contained in a conformationally constrained molecule (Pierschbacher and Ruoslahti, 1987; Ali et al., 1994; Cheng et al., 1994; Gurrath et al., 1992; Baker et al., 1992). A potent, RGD-containing _IIb3 antagonist, SK&F-107260, was generated using a strategy that produced conformational constraints via cyclization (Ali et al., 1994). The conformational and compositional features derived from SK&F-107260 were used in the design of a potent nonpeptide _IIb3 antagonist, SB-207448 (1) (see table 1 for structure). Compound 1 contains a 1,4-benzodiazepine-2-acetic acid nucleus, which mimics the extended C₇ turn conformation of the Gly-Asp moiety in SK&F-107260 (Ku et al., 1993, 1995; Bondinell et al., 1994). Continual optimisation of 1 led to the synthesis of SB-214857, a potent and orally active inhibitor of platelet aggregation (Samanen et al., 1995, 1996). In our study, both SK&F-107260 and SB-214857 were radiolabeled with tritium to high specific activity and were used to characterize direct binding interactions with purified _IIb3. Development of the radioligands allowed us to compare the binding domains of peptide and nonpeptide antagonists on _IIb3.

View this table: [in this window] [in a new window]   TABLE 1 Structure of IIb3 antagonists.

	Experimental Procedures

Top Abstract Introduction Procedures Results Discussion References

Synthesis of [³H]-SK&F-107260 and [³H]-SB-214857. [³H]-SK&F-107260 was prepared with tritium in the N-methyl group of the arginine residue. The key intermediate in this synthesis, N^g-mesitylenesulfonyl-[³H]-methylarginine, was prepared in high specific activity tritium labeled form by treating methyl N-2,4-dimethoxybenzyl-N^g-mesitylenesulfonylarginate with methyl-[³H] iodide and silver oxide (for a review see Landvatter, 1993). Removal of the dimethoxybenzyl group by catalytic transfer hydrogenation (4.4% formic acid in methanol/Pd black) followed by saponification gave N^g-mesitylenesulfonyl-[³H]-methylarginine in 32% overall from methyl-[³H] iodide. Successive solution phase coupling to N²-glycyl-N-2(2-[(4-methylphenyl)methylthio]phenyl)-o-phenylmethyl--aspartamide and (2-carboxyphenyl)-1-ethyldisulfide followed by treatment with` anhydrous hydrogen fluoride gave [³H]-SK&F-107260 at a radiochemical purity of 95.3% after high-performance liquid chromatography purification and a specific activity of 65-85 Ci/mmol.

Purification of _IIb3. Ten units of outdated, washed human platelets (obtained from The American Red Cross, Philadelphia, PA) were lysed by gentle stirring in 3% octylglucoside, 20 mM Tris-HCl, pH 7.4, 140 mM NaCl, 2 mM CaCl₂ at 4°C for 2 hr. The lysate was centrifuged at 100,000×g for 1 hr. The supernatant obtained was split in two equal portions and these were applied to (A) a 5-ml lentil lectin sepharose 4B column (E. Y. Labs, San Mateo, CA) or (B) a 5-ml RGD-affinity column, each preequilibrated with 20 mM Tris-HCl, pH 7.4, 100 mM NaCl, 2 mM CaCl₂, 1% octylglucoside (buffer A). After 2 hr incubation, the column was washed with 50 ml cold buffer A. The lectin-retained glycoproteins were eluted with buffer A containing 10% dextrose. Proteins bound to the RGD affinity column were eluted with buffer A containing 1 mM of Arg-Gly-Asp-Ser. The RGD affinity chromatography procedure resulted in a ~6-fold enrichment in active receptor compared to the lectin-purified material. SDS polyacrylamide gel analysis (Laemmli, 1970) revealed that the _IIb and ₃ subunits constituted >90% of the Coomassie Brilliant Blue-stained protein. Protein determination was based on absorbance at 280 nm and the moles of _IIb3 calculated using a molecular weight of 223.2 kDa. The purified _IIb3 was incorporated in mixed phospholipid vesicles (phosphatidylserine:phosphatidylcholine = 7: 3) as described previously (Parise and Phillips, 1985).

Ligand binding to _IIb3-containing liposomes. Binding assays were performed in a 96-well filtration plate assembly (Millipore Corporation, Bedford, MA) using 0.22-µm hydrophilic durapore membranes. In saturation binding assays, _IIb3-containing liposomes were incubated with various concentrations (0.25-30 nM, diluted in a binding buffer, containing 20 mM Tris-HCl, pH 7.4, 100 mM NaCl and 2 mM CaCl₂) of [³H]-SK&F-107260 (65-86 Ci/mmol) or [³H]-SB-214857 (22 Ci/mmol) at room temperature for 1 hr. For [³H]-SK&F-107260 binding, 0.5 and 0.07 µg of the lentil lectin column and RGD affinity column-purified receptors, respectively was used. As for [³H]-SB-214857 binding, twice the amount of _IIb3 was used. After the incubation, receptor-bound [³H]-SK&F-107260 was separated from the unbound by filtration using a Millipore filtration manifold, followed by washing with ice-cold binding buffer (three times, each 0.2 ml). Bound radioactivity remaining on the filters was determined in 1.5 ml Ready Safe (Beckman, Fullerton, CA) in a Beckman Liquid Scintillation Counter (model LS6800), with 40% efficiency. Nonspecific binding was determined in the presence of 2 µM unlabeled SK&F-107260 or SB-214857. The procedures for ¹²⁵I-echistatin (Amersham Corp., Arlington Heights, IL) binding were identical to those used for the tritiated ligands. Nonspecific binding was consistently less than 1% of the total radioactivity added to the samples. The data presented are representative of five separate receptor preparations or as indicated in the text. All data points are the mean of triplicate determinations. S.D.s of the triplicates were always less than 5% of the mean. Saturation binding data were analyzed by the LIGAND program (Munson and Rodbard, 1980).

Binding of biotinylated-fibrinogen to _IIb3. Biotinylation of fibrinogen and binding of the biotinylated-adhesive protein to _IIb3 were performed essentially as described by Charo et al. (1991).

	Results

Top Abstract Introduction Procedures Results Discussion References

Equilibrium binding of [³H]-SK&F-107260 and [³H]-SB-214857 to _IIb3. The _IIb3 used in the binding studies were purified by RGD affinity column chromatography as described in "Experimental procedures." Specific binding of [³H]-SK&F-107260 to the purified _IIb3 was of high affinity and saturable. Figure 1A shows a typical example of receptor saturation binding with increasing concentrations of [³H]-SK&F-107260 in the absence or in the presence of 2 µM unlabeled SK&F-107260. A Scatchard plot (Scatchard, 1949) of the binding data indicated a single class of binding sites (fig. 1B), exhibiting a dissociation constant (K_d) of 1.19 ± 0.34 nM and a maximum binding capacity (B_max) of 0.24 ± 0.05 mol/mol _IIb3, as determined from five separate experiments.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Equilibrium binding of [3H]-SK&F-107260 to IIb3. A, The RGD affinity purified, liposome-incorporated IIb3 (0.07 µg) was incubated with various concentrations of [3H]-SK&F-107260 for 1 hr at room temperature as described in "Experimental procedure." The amount of total () and specific () radioactivity bound were plotted as a function of the concentration of [3H]-SK&F-107260 added. Nonspecific binding () was determined in the presence of 2 µM unlabeled SK&F-107260. Each point is the mean ± S.E.M. of triplicate determinations. B, The same data was plotted by the method of Scatchard.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Equilibrium binding of [3H]-SB-214857 to IIb3. A, Assays conditions were identical to those of [3H]-SK&F-107260 binding, except that 0.15 µg of RGD affinity purified IIb3 was used. (): total binding; (): specific binding; (): nonspecific binding. B, Scatchard analysis of the binding data in A.

Binding of [³H]-SK&F-107260, [³H]-SB-214857 and ¹²⁵I-echistatin to the lentil lectin purified _IIb3. Previous studies have shown that _IIb3 undergoes conformational changes from low- to high-affinity state during platelet activation (Marguerie et al., 1979). Similar to the receptor on the platelet surface, solubilized _IIb3 exists in at least two different conformations as assessed with specific monoclonal antibodies (Parise et al., 1987; Shattil et al., 1985; Honda et al., 1995). Although fibrinogen and other soluble extracellular matrix proteins bind selectively to activated _IIb3, the small RGD peptides and snake venom peptides, e.g., echistatin, appear to bind to multiple states of the receptor (Plow and Ginsberg, 1988; Kiewiarowski et al., 1994; Gan et al., 1988; Savage et al., 1990). It has been shown that RGD affinity chromatography selectively enriches for _IIb3 that can bind ligands, whereas a lectin column purifies the total receptor population (Konus et al., 1992; Steiner et al., 1989). In our study, we have compared the binding of [³H]-SK&F-107260, [³H]-SB-214857 and ¹²⁵I-echistatin to the RGD-affinity purified and lectin purified _IIb3.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Scatchard analysis of 125I-echistatin binding to lentil lectin-purified IIb3. Assay conditions were identical to those used for [3H]-SK&F-107260 binding as described in "Experimental Procedures."

View this table: [in this window] [in a new window]   TABLE 2 Binding of [3H]-SK&F-107260 and [3H]-SB-214857 to the RGD affinity column purified and the lentil lectin purified IIb3

Reversible binding of [³H]-SK&F-107260 and [³H]-SB-214857 to _IIb3. Figure 4 shows that the binding of [³H]-SK&F-107260 or [³H]-SB-214857 to _IIb3 was completely displaced by 1 µM of either unlabeled SB-214857 or SK&F-107260. Similar results were observed even if the receptor was preequilibrated with [³H]-SK&F-107260 or [³H]-SB-214857 for 1 hr before the unlabeled ligands were added to the incubation. These results demonstrate the reversible nature of [³H]-SK&F-107260 and [³H]-SB-214857 binding to _IIb3.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Reversibility of [3H]-SK&F-107260 and [3H]-SB-214857 binding to IIb3. Binding of 2 nM [3H]-SK&F-107260 or 5 nM [3H]-SB-214857 to IIb3 was determined under various conditions. IIb3 was incubated with the radioligand ligands in the presence or absence of unlabeled SK&F-107260 or SB-214857 for 2 hrs. Alternatively, IIb3 was preequilibrated with the radioligands for 1 hr, 2 µM unlabeled SK&F-107260 or SB-214857 was then added, and the samples were incubated for an additional hr. A total of 0.5 µg and 1 µg of the lentil lectin purified IIb3 was used in the [3H]-SK&F-107260 and [3H]-SB-214857 binding assays, respectively.

Divalent cation dependency. It has been shown that _IIb3 contains multiple cation binding sites and that divalent cations regulate the binding of ligands to _IIb3 (Gulino et al., 1992; Smith et al., 1994). Binding of [³H]-SK&F-107260 and [³H]-SB-214857 to _IIb3 also requires divalent cations. Figure 5 shows that Mg⁺⁺, Ca⁺⁺ and Mn⁺⁺ are able to support the binding of [³H]-SK&F-107260 and [³H]-SB-214857, with Mn⁺⁺ being more effective.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Effects of divalent cations. The lentil lectin purified IIb3 was incubated with 3 nM of [3H]-SK&F-107260 (black bars) or 5 nM of [3H]-SB-214857 (gray bars), in a binding buffer containing 0.1 mM EGTA, in the presence or absence of 1 mM divalent cations (Mg++, Ca++ or Mn++).

Competition of unlabeled ligands for the binding of [³H]-SK&F-107260, [³H]-SB-214857 and biotinylated-fibrinogen to _IIb3. To characterize further the binding of SK&F-107260 and SB-214857 to _IIb3, competition experiments were performed employing a series of peptide and nonpeptide _IIb3 antagonists and using [³H]-SK&F-107260 and [³H]-SB-214857 as radioligands. Figure 6A shows the competition curve of unlabeled SK&F-107260 for the binding of [³H]-SK&F-107260. The dissociation constant (K_i) for SK&F-107260 was determined as 2.2 nM. This value was comparable to the K_d of [³H]-SK&F-107260 obtained in saturation binding experiments shown in figure 1, indicating that the radioactive and nonradioactive forms of SK&F-107260 have similar properties in receptor-binding assays. Also shown in figure 6A are the competition curves of unlabeled SB-214857 and three selected inhibitors (2, 3, 5 and the -dodecapeptide, see table 1 for structures). Compound 2 and 3 are potent nonpeptide _IIb3 antagonists, designed and synthesised by a combination of random screening and structure optimisation (Alig et al., 1992). Peptide 5 (SK&F-106760) is a conformationally constrained cyclic RGD peptide that inhibits platelet aggregation at submicromolar concentrations (Ali et al., 1994). At high concentrations of the inhibitors, binding of [³H]-SK&F-107260 decreased to a plateau, which was less than 5% of total binding. The same set of unlabeled ligands were also active in competing for the binding of [³H]-SB-214857 to _IIb3 (fig. 6B).

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Competition binding assays. The lentil lectin purified IIb3 was incubated with 4.5 nM [3H]-SK&F-107260 (A) or 5 nM [3H]-SB-214857 (B) in the presence of various concentrations of unlabeled SK&F-107260 (), 2 (), 3 (), 5 (), -dodecapeptide () and SB-214857 (). The amount of radioligand bound to IIb3 was determined by rapid filtration as described in "Experimental Procedures." Each data point is the mean of quadruplicate samples.

View this table: [in this window] [in a new window]   TABLE 3 Competition of the binding of [3H]-SK&F-107260, [3H]-SB-214857 and biotinylated-fibrinogen to IIb3 by antagonists: comparison with inhibition of platelet aggregation

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Correlation of affinities values for 13 inhibitors obtained in [3H]-SK&F-107260 and biotinylated-fibrinogen binding assays. (): nonpeptides; (): cyclic RGD peptides; (): linear RGD peptides; () -dodecapeptide; (): echistatin. The slope of the regression line was 0.88, with a correlation coefficient of 0.992.

SB-214857 and -dodecapeptide show competitive inhibition of [³H]-SK&F-107260 binding to _IIb3. To provide additional evidence that SB-214857 and SK&F-107260 share a common binding site on _IIb3, we have performed [³H]-SK&F-107260 saturation binding experiments in the presence of two concentrations of unlabeled SB-214857. As seen in figure 8A, SB-214857 appears to be a competitive inhibitor of [³H]-SK&F-107260 as indicated by its ability to decrease the affinity of [³H]-SK&F-107260 for _IIb3 although not reducing the maximal binding capacity. A similar mode of inhibition was observed when the -dodecapeptide was used, suggesting that it is also a competitive inhibitor of RGD peptide binding to _IIb3 (fig. 8B).

View larger version (24K): [in this window] [in a new window]   Fig. 8.   Scatchard plots of [3H]-SK&F-107260 saturation binding in the presence of antagonists. The lentil lectin purified IIb3 (0.5 µg) was used in the assays. The lines drawn represent the best fit to the data determined by linear regression analysis. The concentrations of [3H]-SK&F-107260 used were 0.3 to 10 nM. A, Binding of [3H]-SK&F-107260 in the absence () or presence of 2 nM () and 4 nM () of unlabeled SB-214857; B, -dodecapeptide: 25 µM (), 50 µM () and 100 µM ().

	Discussion

Top Abstract Introduction Procedures Results Discussion References

We previously reported the design and synthesis of a highly potent 1,4-benzodiazepine-based nonpeptide _IIb3 antagonist, 1, which displays an atom for atom overlay with the conformational and compositional features of SK&F-107260 (Ku et al., 1993). In 1, an arginine mimetic side chain is attached to the benzodiazepine nucleus at 8 position. We also prepared an isomer of 1 in which an arginine mimetic is attached to position 7 of the benzodiazepine (Ku et al., 1995). Further optimisation of structure in the 7-series provided a potent _IIb3 antagonist, SB-214857 (Samanen et al., 1996). Molecular modeling studies suggest it is possible to superimpose the acid, the benzodiazepine nucleus and the arginine mimetic side chain of SB-214857 with those of 1, and with SK&F-107260 (Samanen et al., 1995). Thus, modeling would predict that SB-214857 binds to _IIb3 in the same manner as SK&F-107260. To examine this hypothesis in more detail, our study was performed to directly compare the receptor binding of [³H]-SK&F-107260 and [³H]-SB-214857.

Interactions of [³H]-SK&F-107260 and [³H]-SB-214857 with purified _IIb3 were examined using radioligand binding assays. Two preparations of purified _IIb3 were used. The lentil lectin-purified material contains the total receptor population, although RGD affinity chromatography selectively enriches for receptor capable of binding ligands. Previous studies have shown that _IIb3 can exist in at least two distinct conformational states once solubilized (Parise et al., 1987; Shattil et al., 1985; Honda et al., 1995). We compared the binding of our radioligands using both receptor preparations to investigate their ability to distinguish multiple forms of the receptor. Our studies revealed that [³H]-SK&F-107260 and [³H]-SB-214857 shared many characteristics in binding to _IIb3. Both radioligands bind _IIb3 saturably, reversibly, with high affinity and require divalent cations. The binding is monophasic for each radioligand, indicating a single class of noncooperative binding sites. The B_maxs of binding to two preparations of purified _IIb3, determined by the [³H]-SK&F-107260 and the [³H]-SB-214857 binding assays are identical, suggesting that they bind to the same population of _IIb3. Moreover the binding of our radioligands was comparable to that observed with ¹²⁵I-echistatin (table 2). All three radioligands showed similar B_max values and a single class of binding sites.

Previous work has documented multiple binding sites for divalent cations on _IIb3 (Gulino et al., 1992; Smith et al., 1994; Loftus et al., 1996). Studies with _IIb3 and related integrins (Lampugnani et al., 1991; Suehiro et al., 1996) show that divalent cations can differentially regulate ligand binding selectivity. The very similar effects of Ca⁺⁺, Mg⁺⁺ and Mn⁺⁺ on the binding of [³H]-SK&F-107260 and [³H]-SB-214857 to _IIb3 further support the notion that the radioligands interact with the receptor in a similar manner.

To examine whether the two ligands interact with a similar binding region on _IIb3, competition binding experiments were performed with a number of _IIb3 antagonists using both [³H]-SK&F-107260 and [³H]-SB-214857 as radioligands. These antagonists were selected because of their wide diversity in structures. The K_is determined for these antagonists are very similar in each binding assay (table 3), suggesting that SB 214857 and SK&F-107260 share a common binding site on _IIb3. More importantly, direct binding studies indicated that SB-214857 was a competitive inhibitor of SK&F-107260, causing a shift only in the K_d for [³H]-SK&F-107260 binding to _IIb3 (fig. 8A). The binding of SK&F-107260 and SB-214857 to a similar binding site on _IIb3 suggests that the prerequisite for antagonist activity is a molecule that contains basic and acidic residues spaced appropriately. This is consistent with the wide structural diversity found in nonpeptide _IIb3 antagonists.

Many of the peptides and nonpeptides listed in table 3 have also been characterized for their ability to inhibit aggregation of human platelets activated with 10 µM ADP. The affinity of ligands from the binding results and the rank order of potency for ligands to inhibit aggregation are well correlated. Although the rank order correlates, the absolute potency of ligands in the platelet aggregation assay is apparently lower than the determined binding affinity. This is likely due in large part to the presence of micromolar amounts of endogenous ligands for _IIb3, i.e., fibrinogen, fibronectin, von Willebrands factor and thrombospondin found in plasma and released by platelets. However, the data support the hypothesis that RGD peptides and mimetic ligands inhibit platelet aggregation though competitive inhibition of fibrinogen binding to platelet _IIb3.

We have also directly examined the competition binding of the -dodecapeptide for [³H]-SK&F-107260 to _IIb3. Photoaffinity studies demonstrated that RGD peptides and the -dodecapeptide bind preferentially to the ₃ and _IIb subunits, respectively, and therefore may have distinct binding sites (D'Souza et al., 1988). Our data indicated that the -dodecapeptide is an apparent competitive inhibitor, suggesting that the -dodecapeptide and SK&F-107260 interact with a mutually exclusive, or perhaps a common site on _IIb3.

In summary, we have established radioligand binding assays for _IIb3 using [³H]-SK&F-107260 and [³H]-SB-214857. Previously we proposed that the 1,4-benzodiazepine-based nonpeptide, SB 214857, is a mimetic of cyclic RGD peptide, SK&F 107260, based on compositional and conformational similarities, and now we show that they behave in essentially an identical manner pharmacologically. Moreover, both ligands display pharmacological profiles similar to the binding of the natural ligand, fibrinogen. These results provide evidence that supports the hypothesis that SK&F-107260 and SB-214857 bind to a common site on _IIb3. This binding site is mutually exclusive for the fibrinogen -dodecapeptide suggesting that this ligand binds to a coupled or overlapping site on the fibrinogen receptor. Given the quantitative and reproducible nature of the [³H]-SK&F-107260 and [³H]-SB-214857 binding assays, these radioligands appear to be useful for both pharmacological and mechanistic studies of the _IIb3 integrin.