ROLE OF VITRONECTIN INTERACTION IN THE BIOLOGY OF THE UROKINASE RECEPTOR

Expression of uPAR has been extensively correlated with the malignant progression and metastasis of cancer; however, little evidence for a causal connection between increased uPAR expression and these processes has been documented to date. A complete functional alanine scan of human uPAR, pinpointed the extracellular matrix protein vitronectin (VN) as the critical uPAR-interactor required to induce cell adhesion, migration, and signaling in vitro, identifying this molecular interaction as a possible target for anti-cancer therapy (Madsen et al., 2007). The same study helped to determine the binding epitope in uPAR responsible for its interaction to VN in an integrin-independent fashion. The composite epitope is fully conserved in mouse and included three amino acids (W32, R58, I63) in domain 1 (D1) and 2 amino acids (R92, Y93) in the linker region between D1 and D2. We substituted these residues with alanine by site directed mutagenesis and analyzed the biological activity of resulting receptor variants in CHO Flp-In cells as compared to wild-type muPAR. All the mutant receptors displayed a deficiency for the binding to VN, preserving the receptor binding to its natural ligand uPA. They also failed to induce muPAR-induced cell morphology changes. These changes include the formation of actin-rich lamellipodia, loss of stress fibers, reduced cell-cell contact as well as a complete failure to form colonies when seeded at low density. The mouse uPARW32A (muPARW32A) was chosen for further experiments since it did not show folding problems. Moreover the W32 position is not a part of the receptor chemotactic epitope and is not involved in the receptor cleavage. Additional experiments using recombinant soluble muPAR confirmed the impaired VN-binding and a normal uPA-binding. To determine if uPAR and/or VN are involved in tumor formation and progression, and more specifically, if the direct uPAR/VN-interaction is important in this process, we exploited a xenograft mouse model. For this purpose muPAR or muPARW32A or a muPAR variant lacking of the D1, required for both VN and uPA binding (muPARΔD1), were expressed in HEK293 Flp-In GFP positive cells and injected in the fourth mammary fat pad of immunodeficient mice. In parallel the cells were tested for some in vitro assays. We demonstrated that HEK293 cells expressing muPAR are more proliferating and less apoptotic than the other cells. The interaction with VN is required to increased cell proliferation and to prevent from apoptosis. Moreover, muPAR-VN interaction induced cell spreading and migration. muPAR expressing cells formed palpable tumors earlier than cells expressing the mutant receptors and the tumor growth was significantly faster. Despite the expression of the muPARW32A didn’t affect the timing of the primary tumor formation, it drastically slowed down the growth of the primary tumor mass. Finally we conducted in vitro studies to determine the molecular mechanisms underlying the possible role of the uPAR/VN-interaction in vivo. From these tests emerged that VN may act as an adhesion “bridge” between different cells expressing uPAR and VN-integrins or cells expressing both uPAR, suggesting a possible role of the uPAR/VN-interaction not only in cell-ECM interactions but also in cell-cell adhesion events including the extravasation of metastatic cancer cells.