BACKGROUND Philadelphia cells are human chronic myelogenous leukemia (CML) cells that contain the BCR/ABL oncogene (a fusion of the BCR and ABL genes). Selective eradication of these cells in vitro can be achieved by combined treatment with antisense phosphorothioate oligodeoxynucleotides ([S]ODNs) specifically targeted to this oncogene (bcr/abl [S]ODNs) and a suboptimal (for use as a single agent) dose of mafosfamide (the in vitro active form of cyclophosphamide). OBJECTIVE We evaluated the ability of bcr/abl antisense [S]ODNs, alone or subsequent to treatment with a single injection of cyclophosphamide, to suppress the leukemic process induced in severe combined immunodeficient (SCID) mice by Philadelphia cells (i.e., primary CML-blast crisis [CML-BC] cells). In addition, we studied potential mechanisms that might explain the efficacy of the bcr/abl antisense [S]ODN-mafosfamide combination against Philadelphia cells in vitro. METHODS The effects of treating leukemic mice with cyclophosphamide (25 mg/kg body weight; 25% of the dose required to eradicate evidence of leukemia in SCID mice) and/or bcr/abl antisense [S]ODNs were assessed by analysis of survival, by examination of bone marrow for the presence of leukemia cells (using a colony formation assay or using coupled reverse transcription and the polymerase chain reaction to screen for bcr/abl messenger RNA), and by examination of a variety of tissues for the presence of infiltrating leukemia cells. The induction of apoptosis (a cell death program) in vitro in primary CML-BC cells following treatment with bcr/abl antisense [S]ODNs plus or minus prior treatment with mafosfamide was monitored by use of a commercial assay. Relative cellular uptake of [S]ODNs by CML-BC cells treated in vitro with or without prior treatment with mafosfamide was determined by use of confocal microscopy and flow cytometry (for fluorescent [S]ODNs) or by use of blotting techniques that employed radioactively labeled probes (for extracted, unlabeled [S]ODNs). Levels of specific proteins in treated and untreated cells were determined by use of western blotting methods. Reported P values are two-sided. RESULTS The disease process in leukemic mice was retarded substantially by combination treatment with cyclophosphamide and specific bcr/abl antisense [S]ODNs (P < .001, relative to treatment with specific antisense [S]ODNs alone, cyclophosphamide alone, or cyclophosphamide plus nonspecific [i.e., control] antisense [S]ODNs); 50% of the mice treated with cyclophosphamide and specific antisense [S]ODNs appeared to be cured of leukemia. The combination treatment was associated with increased induction of apoptosis. In addition, cellular uptake of bcr/abl antisense [S]ODNs appeared to be increased twofold to sixfold by prior treatment with mafosfamide. This increased uptake of [S]ODNs was associated with enhanced suppression of p210bcr/abl protein levels. CONCLUSIONS Combination therapy with antisense [S]ODNs targeted to specific oncogenes and less toxic doses of anticancer drugs may represent a rational strategy to purpose for the treatment of human leukemias.