Our earlier studies indicated a role for polyamines (namely, putrescine, spermidine, and spermine) not only in tumor growth but also in tumor metastases. We have observed that administration of alpha-difluoromethylornithine (DFMO), an irreversible inhibitor of ornithine decarboxylase, resulted in significant inhibition of visually detectable pulmonary metastases in mice implanted with Lewis lung carcinoma. The objective of the present study is to investigate the effect of DFMO on other spontaneous and experimental metastatic models and also to determine which step(s) in the tumor metastatic cascade is sensitive to DFMO. The results presented in this study with malignant mouse B16 amelanotic melanoma (B16a) showed a dose-dependent effect of DFMO on the inhibition of both tumor growth and grossly detectable pulmonary metastases. DFMO, when administered as 0.5, 1, and 2% solution in drinking water, resulted in 0, 24.5, and 60% inhibition of tumor growth, respectively, whereas at the same doses an inhibition of 55, 83, and 96% of visible metastases was observed. At treatment levels of 1 and 2% DFMO, 30 and 65% of the animals were free of metastases. DFMO, at 0.5%, did not show any effect on tumor growth, while a significant 55% inhibition of visible pulmonary metastasis was observed, suggesting a specific role for polyamines in tumor metastasis. DFMO treatment also resulted in a significant reduction of putrescine and spermidine levels with a slight increase in spermine concentration in the tumor tissue. DFMO administration did not inhibit the experimental metastases induced as a result of i.v. injection of B16 melanoma (line F10) tumor and Lewis lung carcinoma cells into the tail vein. These results provide preliminary evidence to indicate that tumor cell polyamine depletion by DFMO might affect the first step in the metastatic cascade, intravasation (i.e., prevent the invasion of metastatic tumor cells into lymphatics or blood vessels), although the effect of DFMO on other steps in the metastatic cascade cannot be ruled out.