Recent studies have demonstrated the inhibitory effect of exogenous adenosine on TNF production. During inflammation endogenous adenosine levels are elevated and may be one of several anti-inflammatory mediators that reduce TNF synthesis. In the present study the authors investigated this role of adenosine in freshly isolated human PBMC. The effect of endogenous adenosine on TNF formation was studied by four different approaches. First, adenosine deaminase was added to LPS-stimulated mononuclear cells. This enzyme specifically deaminates extracellular adenosine to the inactive metabolite inosine. TNF production was augmented from baseline stimulation (LPS alone) of 3.5 +/- 0.4 ng ml-1 -5.2 +/- 0.9 ng ml-1 in the presence of 10 U ml-1 adenosine deaminase. Second, TNF production was determined after stimulation in the presence of dipyridamole, an inhibitor of cellular re-uptake of adenosine which increases extracellular concentrations. TNF synthesis was reduced dose-dependently from 3.1 +/- 0.9 ng ml-1 -1.1 +/- 0.2 ng ml-1 by 10 microM dipyridamole. Third, the adenosine A2 receptor antagonist 8-(3-chlorostyryl)caffeine (100 nM) enhanced TNF synthesis from a baseline of 3.7 +/- 0.5 ng ml-1 -5.5 +/- 0.9 ng ml-1. In contrast, no increase resulted from the addition of 100 nM of the specific A1 receptor antagonist 8-cyclopentyl-1,3-dipropylxanthine. Finally, the authors were able to show that suppression of TNF formation by the specific type IV phosphodiesterase inhibitor rolipram can be completely reversed by adenosine deaminase or by the application of the A2 receptor antagonist. The authors conclude that endogenous adenosine controls TNF production. This effect of adenosine may not only have a physiological role but also appears to contribute to the pharmacological inhibition of TNF synthesis by exogenous agents such as the specific type IV phosphodiesterase inhibitor rolipram.