Platelet-adhesive protein-tumor cell interaction was studied in vitro and in vivo. Monoclonal antibody 1OE5, which inhibits binding of fibronectin and von Willebrand factor to the platelet membrane glycoprotein GPIIb-GPIIIa complex, inhibited the binding of mouse CT26 and human HCI8 colon carcinoma cells to platelets by 63-65%, whereas an irrelevant monoclonal antibody, 3B2, had no effect. Monoclonal antibody 6D1, which inhibits binding of von Willebrand factor to GPIb, also had no effect. RGDS, a tetrapeptide that represents the adhesive domain of fibronectin and von Willebrand factor inhibited binding of the tumors to platelets by 64-69%. Monospecific polyclonal antifibronectin antibody inhibited binding by 60-82%; anti-von Willebrand factor antibody inhibited binding by 75-81%.In vivo, polyclonal monospecific anti-mouse von Willebrand factor antibody inhibited pulmonary metastases induced by CT26 tumor cells by 53-64%, B16a amelanotic melanoma cells by 45% and T241 Lewis bladder cells by 46% without induction of thrombocytopenia. Pulmonary metastases with CT26 cells could be inhibited by induction of thrombocytopenia, and reconstituted by infusion of either murine or human platelets. Reconstitution of pulmonary metastases with human platelets could be inhibited 77% by preincubation of human platelets with monoclonal antibody 1OE5 before infusion of platelets into mice.Thus, platelets appear to contribute to metastases by their adhesive interaction with tumor cells via the adhesive proteins fibronectin and von Willebrand factor.

The interactions between ligands containing the recognition sequence arginine-glycine-aspartic acid (RGD) and integrin receptors are important in many cell-cell and cell-protein interactions. The platelet contains five integrin receptors and they contribute significantly to platelet adhesion and aggregation. To investigate the RGD binding domains on platelet integrins, we immobilized a series of RGD peptides containing variable numbers of glycine residues [(G)n-RGDF] on polyacrylonitrile beads and evaluated the ability of the beads to interact with platelets. With native platelets, virtually no interaction occurred with G1-RGDF beads, but the interactions increased as the number of glycine residues increased, plateauing with the G9- RGDF and G11-RGDF beads. ADP pretreatment enhanced the interactions with all of the beads, whereas prostaglandin E1 pretreatment eliminated the interactions with the shortest peptide beads, but only partially inhibited interactions with the longer peptide beads. Monoclonal antibodies to glycoprotein (GP) IIb/IIIa were most effective in inhibiting the interactions, but antibodies to GPIIb/IIIa with similar inhibitory effects on fibrinogen binding varied dramatically in their ability to inhibit the interaction between platelets and immobilized RGD peptides. Our data indicate that the majority of RGD binding sites on GPIIb/IIIa can be reached by peptides that extend out approximately 11 to 32 A from the surface of the bead, and these results are in accord with the dimensions of integrin receptors deduced from electron microscopy. Activation of GPIIb/IIIa facilitates the interactions, but platelet inhibition fails to eliminate the interactions with the longer peptide beads, suggesting that access to the RGD binding site on at least a fraction of the GPIIb/IIIa receptors is always possible for preferred ligands. Finally, we found that the G3-RGDF peptide beads were uniquely sensitive to the activation state of the GPIIb/IIIa receptor.

A thrombin receptor (TR) demonstrating a unique activation mechanism has recently been isolated from a megakaryocytic (Dami) cell line. To further study determinants of peptide ligand-mediated activation phenomenon, we have isolated, cloned, and stably expressed the identical receptor from a human umbilical vein endothelial cell (HUVEC) library. Chinese hamster ovary (CHO) cells expressing a functional TR (CHO-TR), platelets, and HUVECs were then used to specifically characterize a-thrombin-and peptide ligand-induced activation responses using two different antibodies: anti-TR"52 directed against a 20-amino acid peptide spanning the thrombin cleavage site, and anti-TR'" generated against the NH2-terminal 160 amino acids of the TR expressed as a chimeric protein in Escherichia coli. Activation-dependent responses to both athrombin (10 nM) and peptide ligand (20 MM) were studied using fura 2-loaded cells and microspectrofluorimetry.

We studied the defect responsible for Glanzmann thrombasthenia in a patient whose platelets expressed < 5% of the normal amount of GPIIb-IIIa. Genetic and biochemical evidence indicated that the patient's GPIIIa genes were normal. However, DNA analysis revealed the patient homozygous for a G818 A substitution in her GPIIb genes, resulting in a Gly273 -> Asp substitution adjacent to the first GPIIb calcium-binding domain. To determine how this mutation impaired GPIIb-IIIa expression, recombinant GPIIb containing the mutation was coexpressed with GPIIIa in COS-1 cells. The GPIIb mutant formed stable GPIIb-IIIa heterodimers that were not immunoprecipitated by either of two heterodimer-specific monoclonal antibodies, indicating that the mutation disrupted the epitopes for these antibodies. Moreover, the GPIIb in the heterodimers was not cleaved into heavy and light chains, indicating that the heterodimers were not transported from the endoplasmic reticulum to the Golgi complex where GPIIb cleavage occurs, nor were the mutant heterodimers expressed on the cell surface. These studies demonstrate that a Gly273 --Asp mutation in GPIIb does not prevent the assembly of GPIIb-IIIa heterodimers, but alters the conformation of these heterodimers sufficiently to impair their intracellular transport. The impaired GPIIb-IIIa transport is responsible for the thrombasthenia in this patient. (J. Clin. Invest. 1994. 93:172-179.) Key words: bleeding disorder -platelet aggregation * glycoprotein lIb-IIla. in vitro mutagenesis * protein conformation