delta 9-Tetrahydrocannabinol (THC) inhibited nitric oxide (NO) production by mouse peritoneal macrophages activated by bacterial endotoxin lipopolysaccharide (LPS) and interferon-gamma (IFN)-gamma). Inhibition of NO production was noted at THC concentrations as low as 0.5 microgram/mL, and was nearly total at 7 micrograms/mL. Inhibition was greatest if THC was added 1-4 hr before induction of nitric oxide synthase (NOS) by LPS and IFN-gamma, and declined with time after addition of the inducing agents. This suggested that an early step such as NOS gene transcription or NOS synthesis, rather than NOS activity, was affected by THC. Steady-state levels of mRNA for NOS were not affected by THC. In contrast, protein synthesis was inhibited as indicated by immunoblotting. NOS activity was also decreased in the cytosol of cells pretreated with THC. Addition of excess cofactors did not restore activity. Inhibition of NO production was greater at low levels of IFN-gamma, indicating the ability of the cytokine to overcome inhibition. The effectiveness of various THC analogues, in decreasing order of potency, was delta 8-THC > delta 9-THC > cannabidiol > or = 11-OH-THC > cannabinol. The presumably inactive stereoisomer, (+)delta 9-THC, and the endogenous ligand for cannabinoid receptors, anandamide, were weakly inhibitory. Inhibition may be mediated by a process that depends partly on stereoselective receptors and partly on a nonselective process. LPS, IFN-gamma, hormone receptor agonists, and forskolin increased macrophage cyclic AMP levels. THC inhibited this increase, indicating functional cannabinoid receptors. Addition of 8-bromocyclic AMP increased NO 2-fold, and partially restored NO production that had been inhibited by THC. This occurred only under conditions of limited NOS induction, suggesting that the effect of THC on cyclic AMP was responsible for only a small portion of the inhibition of NO.