Cloned BALB/c 3T3 cells transformed in vitro with polyoma virus and maintained in culture (C cells) were compared with respect to their ability to form experimental metastases, with cells from the same clones that were passaged subcutaneously in vivo (CTC cells), thereby gaining a high tumorigenicity phenotype. No correlation was found between a high subcutaneous tumorigenicity potential or the progression state of these cells, and their capacity to form experimental metastases. Both cell types were also tested for their ability to release heparan sulfate degradation products from a naturally produced, sulfate-labeled extracellular matrix (ECM). Whereas the in vitro maintained C cells did not express an endo-beta-D-glucuronidase (heparanase) activity, some of the in vivo passaged CTC cells expressed such an activity. No strict correlation was found, however, between heparanase activity and the ability of CTC cells from individual tumors to form experimental metastases. However, A9 CTC 220 cells which produced a large number of lung metastases also expressed a high heparanase activity. Both heparanase and lung colonization by A9 CTC 220 cells were inhibited to a large extent by heparin, suggesting the involvement of heparanase in the extravasation of these highly metastatic cells. A9 CTC 220 cells were found to release basic fibroblast growth factor (bFGF) from ECM. This release was partially inhibited by carrageenan lambda which also completely inhibited the heparanase activity expressed by these cells. The in vivo acquisition of heparanase activity was not a result of cell fusion between heparanase expressing host-derived cells and heparanase-negative transformed cells. This conclusion was based on the fact that both the in vitro maintained, heparanase-negative as well as the in vitro passaged, heparanase-positive cells exhibited a similar membrane and molecular market profile.