Background: When administered locally, recombinant platelet factor 4 (rPF4), a known angiogenesis inhibitor, has been shown to effectively suppress murine melanoma and human colon carcinoma primary tumor growth in mice. It was tentatively concluded that this effect was due to the inhibition of tumor neovascularization. Purpose: This study has evaluated the effects of systemically administered rPF4 on the growth and establishment of experimental B16F10 melanoma lung metastases in syngeneic mice. Methods: B16F10 cells (0.5–1.0 × 105) were administered intravenously to mice; 21 days later, the lungs were removed and the tumor foci were counted. Treatments with rPF4 were given to C57BL/6J mice immediately following tumor inoculation either (a) intravenously as a single dose (0375, 0.75, 1.5, or 2.0 mg) or as multiple doses (6 mg total) over a 48-hour period, (b) subcutaneously or intramuscularly as multiple doses (6 mg total) over a 72-hour period, or (c) subcutaneously as multiple doses (6 mg total) over a 92-or 96-hour period following a delay of 4, 24, or 48 hours. (BALB/cByJ × C57BL/6)F1 (CByB6F1/J) athymic nude mice received rPF4 subcutaneously over a 72-hour period. The ability of rPF4 to block binding of 5lCr-labeled B16F10 tumor cells to matrix-coated microtiter plates was evaluated in vitro. The in vivo effect of intravenously injected rPF4 on the retention of 51Cr-labeled B16F10 cells was examined by determining the remaining lung-associated radioactivity after 30 minutes or 1, 2, or 4 hours. Results: Intravenous administration of rPF4 significantly inhibited the development of metastatic lung nodules in a dose-dependent fashion as assessed by both lesion number (p<.03) and lung weight (P<.05). When initiation of subcutaneous treatment with rPF4 was delayed 24–48 hours, the number of metastatic foci was significantly reduced (P<.05). The antitumor effect of rPF4 was not dependent on a T-lymphocyte-mediated process, since subcutaneous rPF4 treatment also suppressed the number of lung metastases in T-cell-deficient athymic CByB6F1/J nude mice. In vitro experiments demonstrated modest inhibitory effects (28%) of rPF4 on B16F10 tumor cell binding to purified murine vitronectin. However, lung clearance experiments at early time points (0.5 hour and 1 hour) showed that tumor lodging was not affected by rPF4 treatment. Conclusions: The administration of rPF4 by systemic routes inhibited the development of experimental lung metastases. These findings are consistent with the known angiostatic properties of rPF4, and we conclude that inhibition of metastatic tumor formation may be due to inhibition of tumor-induced neovascularization. Implications: These results support the testing of rPF4 as an angiostatic agent in the treatment of metastatic tumors. [J Natl Cancer Inst 87: 304-309, 1995]