No abstract

Membrane vesicles are shed by tumor cells both in vivo and in vitro. Although their functions are not well understood, it has been proposed that they may play multiple roles in tumor progression. We characterized membrane vesicles from human HT1080 fibrosarcoma cell cultures for the presence of proteinases involved in tumor invasion. By gelatin zymography and Western blotting, these vesicles showed major bands corresponding to the zymogen and active forms of gelatinase B (MMP-9) and gelatinase A (MMP-2) and to the MMP9⅐tissue inhibitor of metalloproteinase 1 complex. Both gelatinases appeared to be associated with the vesicle membrane. HT1080 cell vesicles also showed a strong, plasminogen-dependent fibrinolytic activity in 125 I fibrin assays; this activity was associated with urokinase plasminogen activator, as shown by casein zymography and Western blotting. Urokinase was bound to its high affinity receptor on the vesicle membrane. Addition of plasminogen resulted in activation of the progelatinases associated with the vesicles, indicating a role of the urokinase-plasmin system in MMP-2 and MMP-9 activation. We propose that vesicles shed by tumor cells may provide a large membrane surface for the activation of membrane-associated proteinases involved in extracellular matrix degradation and tissue invasion.

The in vitro release of matrix-degrading proteinases from breast cancer cells is associated in part with shed membrane vesicles. To determine whether shed vesicles might play a similar role in ovarian cancer cells, we analyzed the shedding phenomenon in vivo and in vitro as well as the enzymatic content of their vesicles. This is the first time that an immunoelectron microscopical analysis revealed membrane vesicles carrying tumor-associated antigen alpha-Folate Receptor (alpha-FR), circulating in biological fluids (ascites and serum) of an ovarian carcinoma patient. These vesicles were trapped in a fiber network with characteristic fibrin periodicity. An ovarian cancer cell line (CABA I) established from ascitic fluid cells of this patient, grew in Matrigel and formed tubular structures suggesting invasive capability. Immunofluorescence analysis demonstrated strong cytoplasmic staining of CABA I cells with anti-matrix metalloproteinase-9 (MMP-9) and anti-urokinase-type plasminogen activator (uPA) antibodies. CABA I cells shed membrane vesicles, which were morphologically similar to those identified in vivo, as determined by electron microscopy. Gelatin zymography of vesicles isolated both in vivo and in vitro revealed major gelatinolytic bands of the MMP family, identified as the zymogen and active forms of gelatinase B (MMP-9) and gelatinase A (MMP-2). By casein-plasminogen zymography we observed high-molecular weight (HMW)-uPA and plasmin bands. Incubation of purified vesicles from CABA I cells with Matrigel led to cleavage of Matrigel components. Taken together, our results point to a possible role of shed vesicles, both in vivo and in vitro, in proteolysis that mediates invasion and spread of ovarian epithelial carcinoma cells.

Abstract. In order to investigate the mechanism by which oligodendrogliomas cause neuronal damage, media conditioned by G26/24 oligodendroglioma cells, were fractionated into shed vesicles and vesicle-free supernatants, and added to primary cultures of rat fetal cortical neurons. After one night treatment with vesicles, a reproducible, dose-dependent, inhibitory effect on neurite outgrowth was already induced and, after 48-72 h of incubation, neuronal apoptosis was evident. Vesicle-free supernatants and vesicles shed by NIH-3T3 cells had no inhibitory effects on neurons. Western blot analyses showed that treated neurons expressed a decreased amount of neurofilament (NF), growth-associated protein (GAP-43) and microtubule-associated protein (MAP-2). Moreover procaspase-3 and -8 were activated while Bcl-2 expression was reduced. Vesicles were found positive for the proapoptotic molecule, Fas-ligand (Fas-L), and for the B isoform of Nogo protein, a myelin component with inhibitory effects on neurons. Nogo B involvement in the vesicle effects was analyzed both by testing the neutralizing capability of anti-Nogo antibodies and by removing the Nogo receptor from neurons by phospholipase C digestion. These treatments did not revert the vesicle effects. To test the role of Fas-L, vesicles were treated with functional anti-Fas-L monoclonals. Vesicle inhibitory and proapoptotic effects were reduced. Vesicles shed by ovarian carcinoma cells (OvCa), which are known to vehicle biologically active Fas-L, had similar effects on neurons to those of oligodendroglioma vesicles, and their inhibitory effects were also reduced by anti Fas-L antibodies. We therefore conclude that vesicles shed by G26/24 cells induce neuronal apoptosis at least partially by a Fas-L mediated mechanism.