We evaluated the effects of natural purine and pyrimidine nucleosides on protection from or reversal of 3'-azido-3'-deoxythymidine (AZT) cytotoxicity in human bone marrow progenitor cells by using clonogenic assays. The selectivity of the "protection" or "rescue" agents was examined in evaluating the antiretroviral activity of AZT in combination with these modulating agents and of AZT alone. Following exposure of human granulocyte-macrophage progenitor cells for 2 h to 5 ,uM AZT (70% inhibitory concentration), increasing concentrations of potential rescue agents were added. Cells were cultured, and colony formation was assessed after 14 days. At concentrations of up to 50 ,uM no natural 2'-deoxynucleosides, including thymidine, were able to reverse the toxic effects of AZT. Dose-dependent reversal was observed with uridine and cytidine, and essentially complete reversal was achieved with 50 ,uM uridine. In the protection studies, 100 ,uM thymidine almost completely antagonized the inhibition of granulocyte-macrophage colony formation produced by 1 ,zM AZT (50% inhibitory concentration), and 50 ,uM uridine effected 60% protection against a toxic concentration of AZT (5 ,uM) (70% inhibitory concentration). The antiretroviral activity of AZT in human peripheral blood mononuclear cells, assessed by reverse transcriptase assays, was substantially decreased in the presence of thymidine, whereas no impairment of suppression of viral replication was observed in the presence of uridine in combination with AZT at a molar ratio (uridine/AZT) as high as 10,000. This demonstration of the capacity of uridine to selectively rescue human bone marrow progenitor cells from the cytotoxicity of AZT suggests that use of uridine rescue regimen with AZT may have potential therapeutic benefit in the treatment of acquired immunodeficiency syndrome.3'-Azido-3'-deoxythymidine (AZT), a pyrimidine nucleoside synthesized two decades ago by Horwitz et al. (7), has recently been shown to transiently improve certain immunological functions in some patients with acquired immunodeficiency syndrome (AIDS) (3), resulting in a decrease in the incidence of opportunistic infections and prolonging survival. The antiretroviral effects of AZT are probably based upon its conversion through cellular kinases to AZT triphosphate, which binds to reverse transcriptase and thereby inhibits viral DNA synthesis by chain termination (4). Although AZT selectively inhibits the replication of human immunodeficiency virus type 1 (HIV) (10), its applications in preliminary clinical trials (11,15) were limited by expressions of bone marrow toxicity. Consistent with these expressions we recently reported (12) that continuous exposure to AZT for 14 days effected a dose-dependent inhibition of human granulocyte-macrophage CFU (CFU-GM) and erythroid burst-forming unit colonies, the 50% inhibitory concentrations being 0.9 ± 0.1 and 2.4 ± 0.4 ,uM for the respective colonies. Several pharmacologic approaches are potentially available to improve the chemotherapeutic selectivity...