Although cytosine arabinoside (araC) can induce a remission in a majority of patients presenting with acute myeloblastic leukemia (AML), a minority fail to respond and moreover the drug has less effect in acute lymphoblastic leukemia (ALL). The carrier-mediated influx of araC into purified blasts from patients with AML, ALL, and acute undifferentiated leukemia (AUL) has been compared to that of normal lymphocytes and polymorphs. Blasts showed a larger mediated influx of araC than mature cells, since mean influxes for myeloblasts and lymphoblasts were 6- and 2.3-fold greater than polymorphs and lymphocytes, respectively. Also, the mean influx for myeloblasts was fourfold greater than the mean for lymphoblasts. The number of nucleoside transport sites was estimated for each cell type by measuring the equilibrium binding of [(3)H]nitrobenzylthioinosine (NBMPR), which inhibits nucleoside fluxes by binding with high affinity to specific sites on the transport mechanism. The mean binding site numbers for myeloblasts and lymphoblasts were 5- and 2.8-fold greater, respectively, than for the mature cells of the same maturation series. The mean number of NBMPR binding sites for myeloblasts was fourfold greater than for lymphoblasts. Patients with AUL were heterogeneous since blasts from some gave values within the myeloblastic range and others within the lymphoblastic range. The araC influx correlated closely with the number of NBMPR binding sites measured in the same cells on the same day. Transport parameters were measured on blasts from 15 patients with AML or AUL who were then treated with standard induction therapy containing araC. Eight patients entered complete remission, while seven failed therapy, among whom were the three patients with the lowest araC influx (<0.4 pmol/10(7) cells per min) and NBMPR binding (<3,000 sites/cell) for the treated group. In summary, myeloblasts have both higher araC transport rates and more nucleoside transport sites than lymphoblasts and this factor may contribute to the greater sensitivity of AML to this drug. AraC transport varied >10-fold between leukemic blasts and normal leukocytes, but transport capacity related directly to the number of nucleoside transport sites on the cell. Finally, low araC transport rates or few NBMPR binding sites on blasts were observed in a subset of patients with acute leukemia who failed to achieve remission with drug combinations containing araC.