Both peripheral blood monocytes and neutrophils are known to be capable of lysing a variety of extracellular tumor and non-tumor cell targets. The HL-60 human promyelocytic leukemia cell line has served as a useful model of human granulocyte and macrophage differentiation in studies from many laboratories. We have previously reported that phorbol diesters, which induce differentiation along the macrophage pathway, stimulate HL-60 cells to become strikingly cytotoxic to a variety of red cell targets. We now report that agents known to differentiate HL-60 along the granulocyte pathway (retinoic acid, dimethyl sulfoxide, 5-azacytidine) do not, in themselves, induce HL-60 to become cytotoxic. However, previous exposure (3-5 days) to these granulocyte pathway active agents markedly enhances phorbol diester-triggered killing. This enhancement is particularly striking at decreased effector:target ratios (as low as one effector per five targets) and is also demonstrated by a shift to lower concentrations of the phorbol diester dose-response curve. Retinoic acid is the most effective of the three agents tested, although priming (previous exposure) with dimethyl sulfoxide or 5-azacytidine also markedly enhances killing. These studies demonstrate that HL-60-mediated killing may be dissected pharmacologically into at least two distinct steps and further support the utility of this model system in studies of the development of macrophage-like cytotoxic cells. This system has also proven to be useful in the characterization of cytokines which mimic the differentiation effects of retinoic acid and dimethyl sulfoxide (J. A. Leftwich and R. E. Hall, manuscript in preparation).