The topoisomerase (topo) II-directed agents etoposide, daunorubicin (DNR), and amsacrine (m-AMSA) are widely used in the treatment of acute myelogenous leukemia (AML). In the present study, multiple aspects of topo II-mediated drug action were examined in marrows from adult AML patients. Colony-forming assays revealed that the dose of etoposide, DNR, or m-AMSA required to diminish leukemic colony formation by 90% (LD90) varied over a greater than 20-fold range between different pretreatment marrows. Measurement of nuclear DNR accumulation in the absence and presence of quinidine revealed evidence of P-glycoprotein (Pgp) function in 8 of 82 samples at diagnosis and 5 of 36 samples at first relapse, but the largest quinidine-induced increment in DNR accumulation (< 2-fold) was too small to explain the variations in drug sensitivity. Restriction enzyme-based assays and sequencing of partial topo II alpha and topo II beta cDNAs from the most highly resistant specimens failed to demonstrate topo II gene mutations that could account for resistance. Western blotting of marrow samples containing greater than 80% blasts revealed that the content of the two topo II isoenzymes varied over a greater than 20-fold range, but did not correlate with drug sensitivity in vitro or in vivo. In addition, levels of topo II alpha and topo II beta in 46 of 47 clinical samples were lower than in human AML cell lines. Immunoperoxidase staining showed that these low topo II levels were accompanied by marked cell-to- cell heterogeneity, with topo II alpha being abundant in some blasts and diminished or absent from others. There was a trend toward increasing percentages of topo II alpha-positive cells in pretreatment marrows that contained more S-phase cells. Consistent with this observation, treatment of patients with granulocyte-macrophage colony- stimulating factor for 3 days before chemotherapy resulted in increases in topo II alpha-positive cells concomitant with increases in the number of cells traversing the cell cycle. These observations have implications for the regulation of topo II in AML, for the use of topo II-directed chemotherapy, and for future attempts to relate drug sensitivity to topo II levels in clinical material.