Macrophages are important target cells for human immunodeficiency virus type 1 (HIV-1) infection. We have developed a drug targeting system for the selective delivery of phosphorylated nucleoside analogues to these phagocytosing cells. This system is based on the possibility of encapsulating the phosphorylated drugs into autologous erythrocytes and on the subsequent selective modification of their membranes to promote macrophage recognition and phagocytosis. Targeted delivery of phosphorylated nucleoside analogues to human, feline, and murine macrophages inhibits the infectivity of HIV-1, feline immunodeficiency virus, and LP-BM5 viruses more efficiently than the administration of the corresponding nucleoside analogues. In vivo administration of 2',3'-dideoxycytidine 5'-triphosphate (ddCTP) encapsulated into autologous erythrocytes to LP-BM5-infected mice was found to reduce infectivity and disease progression. Furthermore, the simultaneous administration of AZT or ddC produced additive antiviral effects. The possibility of using red cells as drug targeting systems was useful for the design, synthesis, and delivery of new antiviral nucleoside analogues. As a prototype of these new drugs, di-(thymidine-3'-azido-2',3'-dideoxy-D-riboside)-5'-5'-p1-p2-pyrophospha te (AZTp2AZT) was prepared. Although this drug in solution has the same antiviral activity as AZT, when administered encapsulated into erythrocytes it was several times more efficient in inhibiting the infectivity of human, feline, and murine immunodeficiency viruses. Thus, the availability of a drug targeting system for the selective delivery of antivirals to macrophages offers an additional possibility for the development of new drugs and of new combination antiviral therapies.