Interferons (IFNs) have established antitumor action; the mechanism underlying this effect is, however, not yet clear. To probe the possible contribution of inhibition of angiogenesis, we have assessed angiogenesis in the mouse initiated by either human or murine tumor cell lines. Whether test cells were inoculated in the dermis or tumor fragments were grafted onto the cornea, tumor-induced angiogenesis (TIA) was inhibited by IFNs. TIA was also inhibited by the potent IFN inducer polyriboinosinic-polyribocytidylic acid. The effect of IFN was species specific; human IFNs inhibited human tumors and mouse IFNs inhibited murine tumors. This effect suggested that in contrast to other angiogenesis inhibitors, IFNs modulated the signal for angiogenesis produced by the tumor cells. Tumor cells treated in vitro with homologous IFN were significantly (P less than 0.005) less competent to initiate angiogenesis than were untreated cells. Inhibition of angiogenesis was achieved whether vascular response was assessed 1 or 3 days after tumor cell inoculation, suggesting that antiangiogenesis activity was independent of the antiproliferative effects of IFNs. To further substantiate this, L1210 leukemia cells, resistant to the antiproliferative effects of IFNs, were treated with 500 units/ml IFN-beta. IFN had no effect on their proliferation, but in four separate experiments, L1210R cells were impaired in their ability to induce angiogenesis. Thus, inhibition of TIA by IFNs was species specific, occurred at least partly by modulation of the signal inducing angiogenesis, and was expressed in the absence of antiproliferative effects. IFNs also inhibited immunologically induced angiogenesis, whether initiated by allogeneic lymphocytes (LIA) or by the mouse's own T-cells in response to an exogenous antigen (sheep RBC). LIA was markedly suppressed by treatment of host mice with homologous IFN-beta. For example, mean vessel counts induced by allogeneic mouse lymphocytes were decreased from 22.8 +/- 1.4 (SE) to 12.5 +/- 0.8 (P less than 0.0001); mouse IFN-beta had no corresponding effect on xenogeneic human lymphocytes (mean vessel counts decreased to 21.7 +/- 2.6 from 22.7 +/- 2.0). Treatment with human IFN-alpha, -beta, or -gamma in vitro or host mice in vivo reduced the ability of inoculated human peripheral blood lymphocytes to initiate xenogeneic LIA. Inhibition of LIA required a lower dose and/or a shorter incubation period than that needed to modulate TIA. Treatment of the donor of the allogeneic spleen cells in vivo with murine IFN or inducers also resulted in lesser LIA.(ABSTRACT TRUNCATED AT 400 WORDS)