4-[2-(2-Amino-4-oxo-4,6,7,8-tetrahydro-3H-pyrimidino[5,4,6][1,4] thiazin-6-yl)-(S)-ethyl]-2,5-thienoylamino-L: -glutamic acid (AG2034), is a classical antifolate, an analog of folic acid that has been shown to be an excellent inhibitor of glycinamide ribonucleotide formyltransferase (GARFT), ultimately inhibiting the de novo synthesis of purines. We examined the effect of this drug on cell proliferation, steady-state ATP levels, de novo and hypoxanthine salvage ATP synthesis, and on the phosphorylation of AMP kinase, in two different androgen independent prostate cancer cell lines, DU145 and PC-3. Cells were maintained in culture medium containing 10 nM 5-methyl tetrahydrofolate supplemented with or without 1.7 microM hypoxanthine and 1.5 microM thymidine. Cytotoxicity of AG2034 was determined by clonogenic assays. AG2034-induced inhibition of cell proliferation was determined by electronic counting of cells over varying periods of time. Total cellular AMP and ATP pre- and post-drug treatment was quantified by reverse-phase HPLC. [(14)C]-Glycine incorporation and [(3)H]-hypoxanthine conversion into ATP were determined by liquid scintillation counting of HPLC isolated ATP fractions. The phosphorylation of AMP kinase (AMPK) was detected by western blotting. In the absence of 1.7 muM hypoxanthine, AG2034 was cytotoxic to both DU145 and PC-3 cells. In its presence, the cells remained cytostatic for 14 days after which time DU145 but not PC-3 re-initiated growth that was maintained for 35 days even though steady-state levels of ATP in both cell lines remained depleted and [(14)C]-glycine incorporation into ATP was inhibited by >95%. Salvage purine synthesis as measured by incorporation of [(3)H]-hypoxanthine into ATP was maintained in both cell lines albeit to different levels. When AG2034 was added to the culture medium in the presence or absence of 1.7 microM hypoxanthine, cellular ATP levels were reduced by 80% within 24 h in both the cell lines. In the absence of hypoxanthine, the AMP/ATP ratio in PC-3 cells increased by 38% and was accompanied by a modest increase in the level of phosphorylated AMPK; no increase was observed in the presence of hypoxanthine where the AMP/ATP ratio increased by approximately 10%. Under these same culture conditions, the AMP/ATP ratio in DU145 cells in the absence of hypoxanthine increased by 60% and was accompanied by a large increase in phosphorylated AMPK. In the presence of hypoxanthine however, even though the AMP/ATP ratio increased 2.5-fold, phosphorylated AMPK levels did not increase. The cytostatic versus the cytotoxic effect of AG2034 on PC-3 and DU145 cells is mediated by the presence or absence, respectively, of physiological levels of hypoxanthine (1.7 muM) in the media. The ability of DU145 as opposed to PC-3 cells to proliferate in the presence of AG2034 is independent of the intracellular concentration of ATP. Activation of the AMPK signaling pathway in drug-treated PC-3 and DU145 cells is cell line dependent and independent of the AMP/ATP ratio.