Several anti-HIV drugs acting on different steps of virus replication were tested in our experimental model of primary monocyte/macrophages; the results were compared with the activity found in lymphocytes. Nucleoside analogues (AZT, ddI, ddC, d4T, PMEA, 3TC etc.) show greater activity in macrophages (M/M) than in lymphocytes. In particular, the EC50 of AZT, ddC, and ddI in M/M is 2- to 100-fold lower than that found in lymphocytes. This greater efficacy of nucleoside analogues in M/M depends on the enhancement of their chain-terminating activity by the low levels of endogenous deoxynucleoside-triphosphates (dNTP) usually found in resting cells such as M/M. Non-nucleoside reverse transcriptase inhibitors (NNRTI) do not act as chain terminators (thus their antiviral effect is not related to the intracellular concentrations of dNTP); as a consequence the activity of TSAO, HEPT, TIBO, and other NNRTI tested in M/M is similar to that found in lymphocytes. Regarding inhibitors of binding and fusion of HIV, we found that their anti-HIV activity is markedly decreased (or even nullified) when M/M are treated with cytokine activators of M/M function and enhancers of HIV replication. More relevant from a clinical standpoint, protease inhibitors are able to inhibit HIV replication in chronically infected macrophages (i.e., cells carrying the proviral genome already integrated in the host genome). All other inhibitors of late stage of virus life cycle tested (antisense-rev, anti-tat, interferon-alpha and -gamma, phosphorothioate analogues, GLQ-223, etc.) were totally inactive in chronically infected macrophages. The different effects of various classes of HIV inhibitors in lymphocytes and macrophages suggests that AIDS therapy should consider all aspects of the pathogenesis of HIV infection and must be restricted to drugs, or combinations of drugs, active against both lymphocytes and M/M in all body compartments where the virus hides and replicates.