Recent studies report a key role for tumour-derived extracellular vesicles (ECVs, including exosomes and microvesicles) in tumour progression. While ECV functions have been demonstrated for melanoma and breast cancer, similar findings have not been reported for prostate cancer-derived ECVs. We have previously characterized the PC3 prostate cancer cell ECV proteome, reporting a quantitative reduction in one third of the proteins upon cavin-1 expression1 concomitant with reduced xenograft tumour growth and metastasis 2. Here, we show that PC3 derived ECVs migrate to bone and lung when injected into NOD/SCID mice. Cavin-1 expression attenuated PC3 ECV function on osteoclastogenesis and osteoblast proliferation in vitro. Cavin-1 expression ablated ECV-driven osteoclastogenesis while uptake into pre-osteoclasts was attenuated, suggesting loss of a key population of ECV, or of key osteoclastogenic molecules from the same population of ECVs. Electron microscopic assessment of ECVs from cavin-1 and control PC3 cells showed little difference in morphology and size distribution. To further probe the mechanism of cavin-1-mediated ECV protein reduction, we evaluated the relative contributions of the exosome and microvesicle pathways. A slight but consistent increase in the exsome marker CD63 and cargo protein cofilin-1 ECVs derived from cavin-1-PC3 cells. On the other hand, several known plasma membrane proteins were decreased in cavin-1-PC3 ECVs. We confirmed the subcellular localization of two such proteins, EphA2 and 4F2 (CD98hc/SLC3A2), and further showed that cavin-1 expression caused their intracellular translocation, reminiscent of the cavin-1 induced cholesterol re-distribution1. To investigate the involvement of cholesterol-dependent membrane microdomains (lipid rafts), biochemical fractionation and subcellular colocalization experiments were performed. Both EphA2 and 4F2 fractionated in the detergent resistant membranes (DRMs) and co-localized with cholesterol as labell1 ed by filipin. In contrast, neither EphA2 or 4F2 co-localized with CD63, which was not detectable in DRMs by western blotting. Taken together, our data suggest that cavin-1 expression indirectly altered ECV protein content by modulating lipid raft-dependent trafficking to/from the plasma membrane.