Lycopene has been shown to inhibit tumor metastasis in vitro, but it is unclear whether lycopene is antimetastatic in vivo. Here, nude mice were orally supplemented 2 times per week for 12 wk with a low or high dose of lycopene [1 or 20 mg/kg body weight (BW)] or with beta-carotene (20 mg/kg BW). Two weeks after the beginning of supplementation, mice were injected once with human hepatoma SK-Hep-1 cells via the tail vein. Plasma levels of matrix metalloproteinase (MMP)-2 and vascular endothelial growth factor (VEGF) increased gradually in tumor-injected mice (tumor controls) following tumor injection but were markedly lowered by lycopene or beta-carotene supplementation. Ten weeks after tumor injection, mice were killed and tumor metastasis was found to be confined to the lungs. Compared with the tumor controls, high-lycopene supplementation lowered the mean number of tumors from 14 +/- 8 to 3 +/- 5 (P < 0.05) and decreased tumor cross-sectional areas by 62% (P < 0.05). High-lycopene supplementation also decreased the positive rate of proliferating cellular nuclear antigen (PCNA), the level of VEGF, and protein expressions of PCNA, MMP-9, and VEGF in lung tissues. However, high-lycopene increased the protein expression of nm23-H1 (an antimetastatic gene) by 133% (P < 0.001). For most variables measured, effects of lycopene were dose dependent and the effect of beta-carotene was between those of high-dose and low-dose lycopene. These results show that lycopene supplementation reduces experimental tumor metastasis in vivo and suggest that such an action is associated with attenuation of tumor invasion, proliferation, and angiogenesis.