Growth of numerous malignant tumors depends on an exogenous methionine (MET) supply, while endogenously synthesized MET supports normal cell proliferation. Because an antitumor effect should be obtained by aggravating the altered MET metabolism in gliomas, MET dependency of human xenografted gliomas was evaluated and a therapeutic approach using MET deprivation or MET analogs to induce MET starvation was applied. In vitro proliferation inhibition of glioma cell lines by MET deprivation and two MET analogs, ethionine (ETH) and trifluoromethylhomocysteine (TFH), was measured. Proliferation of 7 human glioma cell lines tested was inhibited in MET-free medium, and was poorly or not reversed by homocysteine (HCY). ETH or TFH (concentration range: 0.005-2 mg/ml) inhibited proliferation of all cell lines tested. MET analog-induced inhibition was abolished by MET and enhanced by HCY. Cell-cycle alterations due to MET deprivation were optimally assessed after 30 h of culture and bromodeoxyuridine incorporation. In MET- medium, cells were arrested in the G1-phase. ETH induced a dramatic accumulation of cells in the G2-phase. ATP contents were reduced by MET analogs only in HCY+ medium, suggesting complementary effects of MET analogs and HCY. Human glioma bearing nude mice were fed an amino acid-substituted MET- HCY-supplemented diet (MET-HCY+) and/or treated with MET analogs, injected intraperitoneally daily. Using two human xenografted tumors derived from gliomas, antitumor effects were obtained by subjecting tumor-bearing nude mice to MET starvation. TG-1-MA was more sensitive to MET depletion (40% of growth inhibition, P < 0.10) than TG-8-OZ (no growth inhibition). Antitumor effects of a MET-HCY+ diet and 200 mg/kg of ETH were potentiated when co-administered to glioma-bearing mice (77% GI, P < 0.025 and 67%, P < 0.0057 to TG-1-MA and TG-8-OZ respectively). A dose-response effect with no toxicity was obtained when the ETH dose was increased 10 fold. Potentiation of the effects of ETH and a MET-free diet indicates that they probably act on the same pathway but not the same target. In conclusion, experimentally induced MET deprivation and MET-analog treatment retarded the growth of human gliomas. Combination of MET-analog therapy with MET substitution by HCY enhanced their respective effects.