Microvesicles (diameter ca 200 nm) from the cell-free supernatant of U87MG human glioblastoma cell caused platelet aggregation and coagulation in a manner identical with that previously shown for the intact cells. Both activities were inhibited by dansylarginine -N-(3- ethyl-1,5-pentanediyl) amide (DAPA), confirming the thrombin-dependent nature of both activities. The specific activities per microgram of protein were 2–10 times greater in the microvesicles than in the plasma membrane fraction, suggesting localization in specific membrane domains. Sucrose density centrifugation gave a single protein peak (density 1.14) with congruent procoagulant and platelet aggregating activities. Both activities required the extrinsic pathway, as shown by studies with factor-deficient plasmas, and both were inhibited by heating (60 min/100 degrees C), by reduction and alkylation, and by incubation of the microvesicles with rabbit anti-bovine brain tissue factor antibody. These observations were confirmed using microvesicles from the HL-60 human promyelocytic leukemia cells, which are known to contain tissue factor activity. The results suggest that both procoagulant and proaggregating activities are causally related through the presence of tissue factor in the microvesicles. Studies with the Baumgartner perfusion apparatus showed that U87MG microvesicles increased the size of adherent thrombi nearly tenfold and that these thrombi were associated with nucleated cells from the blood. The increase in adherent thrombi did not occur if perfusion was carried out in the presence of DAPA, confirming the role of thrombin in their formation.

The use of the Baumgartner perfusion system allows the morphometric quantification of platelets interacting with vessel wall, however it presents the basic difficulties of morphometrical measurements. In order to facilitate the procedure of evaluation we developed a semiautomated method to avoid the complexity of the classical evaluation. Our system consists on an optical picture analysis system connected with a specially developed computer program which allows fast quantification. Simultaneously to the outlining of interacting platelets the computer program recognizes, corrects, selects and stores the information, in order to perform the final calculations as previously established. This system has been demonstrated to be as effective as the classical morphometric evaluation in the measure of platelets interacting with subendothelium. Potential sources of error such as subjectivity of the observers in selecting the class of interacting platelets are avoided. The use of this combined method opens the possibility to adapt the Baumgartner perfusion system to clinical routine and to the screening of drugs that modify platelet adherence.

To characterize the platelet receptor sites and the platelet metabolic pathways involved in tumor-cell-induced platelet aggregation, we have used a homologous system consisting of human platelets and 2 tumor cell lines of human origin, which activate platelets through different mechanisms. Preincubation of platelets with an MAb against platelet glycoprotein Ib partially blocked tumor-cell-induced platelet aggregation, and preincubation of platelets with an MAb against the glycoprotein complex GPIIb/IIIa totally blocked the aggregation induced by the 2 tumor-cell lines. No inhibitory effect was found when platelets were treated with PAF-receptor antagonists or with specific peptides which block the platelet sites involved in bacterially induced platelet aggregation. Compounds which raised intra-platelet cAMP levels inhibited tumor-cell-induced platelet aggregation in a dose-related manner. Inhibition of cyclo-oxygenase by aspirin which blocked TxB2 formation by platelets did not inhibit platelet aggregation induced by tumor cells whereas the BW755 compound which inhibits cyclo- and lipoxygenase blocked platelet aggregation. These results demonstrate that tumor-cell-induced platelet aggregation is a glycoprotein-dependent and a lipoxygenase-associated phenomenon.

Platelets are thought to be involved in the development of blood borne metastasis. Ultrastructural and experimental studies demonstrate that association between tumor cells and platelets with subsequent activation of the coagulation cascade takes place in malignancy. Several hypotheses have been proposed to explain the mechanisms by which tumor cells activate platelets including generation of thrombin, ADP release and involvement of arachidonate metabolism. Perfusion studies with human homologous systems showed that intact tumor cells and tumor cell microvesicles were able to induce platelet thrombogenicity under defined flow conditions. The presence of divalent cations and plasma factors was necessary for the cancer cells to exert their activating capacity. These results suggest a role for platelets in the development of secondary metastasis as well as in the thrombotic events of malignancy.

The effects of human recombinant interleukin-1 alpha and beta (rIL-1 alpha; rIL-1 beta) on the adhesion of human A549 lung carcinoma cells and M6 melanoma cells (TC) to human endothelial cells (HECs) in vitro were studied, and on TC/lung entrapment in vivo. In vitro, there was a significant increase in TC/HEC adhesion to HECs pretreated for 4 h with rIL-1 alpha or rIL-1 beta. The effects of rIL-1 alpha and beta on TC/HEC adhesion were time dependent and reached a plateau within 4-6 h. TC/HEC adhesion was not blocked when measured in the presence of antibodies to either fibronectin, glycoprotein IIb/IIIa, anti-ICAM, or anti-LFA. However, enhanced TC/HEC adhesion was completely blocked in the presence of the peptide, GRGDS. In vivo, pretreatment of nude mice for 4 h with rIL-1 alpha (given i.p. before i.v. injection of TCs) enhanced TC retention in the lung 24 h later. Our data demonstrate that IL-1 enhances TC adhesion to the vascular surface both in vitro and in vivo, suggesting that IL-1 can facilitate the metastatic process.