Activated alpha5beta1 integrin occurs specifically on tumor cells and on endothelial cells of tumor-associated vasculature, and plays a key role in invasion and metastasis. The PHSCN peptide (Ac-PHSCN-NH(2)) preferentially binds activated alpha5beta1, to block invasion in vitro, and inhibit growth, metastasis and tumor recurrence in preclinical models of prostate cancer. In Phase I clinical trial, systemic Ac-PHSCN-NH(2) monotherapy was well tolerated, and metastatic disease progression was prevented for 4-14 months in one-third of treated patients. We have developed a significantly more potent derivative, the PHSCN-polylysine dendrimer (Ac-PHSCNGGK-MAP). Using in vitro invasion assays with naturally serum-free basement membranes, we observed that the PHSCN dendrimer was 130- to 1900-fold more potent than the PHSCN peptide at blocking alpha5beta1-mediated invasion by DU 145 and PC-3 human prostate cancer cells, whether invasion was induced by serum, or by the Ac-PHSRN-NH(2) peptide, under serum-free conditions. The PHSCN dendrimer was also approximately 800 times more effective than PHSCN peptide at preventing DU 145 and PC-3 extravasation in the lungs of athymic mice. Chou-Talalay analysis suggested that inhibition of both invasion in vitro and extravasation in vivo by the PHSCN dendrimer are highly synergistic. We found that many extravasated DU 145 and PC-3 cells go onto develop into metastatic colonies, and that a single pretreatment with the PHSCN dendrimer was 100-fold more affective than the PHSCN peptide at reducing lung colony formation. Since many patients newly diagnosed with prostate cancer already have locally advanced or metastatic disease, the availability of a well-tolerated, nontoxic systemic therapy, like the PHSCN dendrimer, which prevents metastatic progression by inhibiting invasion, could be very beneficial.