Since extracellular adenosine can alter adenylate cyclase activity and cell function in a variety of tissues, the effect of adenosine on mediator release was studied with isolated rat mast cells. Adenosine did not alter the rate of spontaneous release, but histamine release induced by anti-IgE, concanavalin A (Con A), 48/80, or the calcium ionophore A23187 was markedly enhanced by 1 nM to 0.1 mM adenosine. Mast cells preincubated in the presence of 10 µM adenosine and then challenged with 0.1 µg/ml of A23187 released 71.5% (± 7.4 SEM) of the cellular histamine compared to a 38.2% (± 3.0 SEM) release from control cells. Adenosine caused a 3.0-fold (± 0.2 SEM) increase in the rate of release occurring 30 to 60 sec after A23187 addition, but did not change the duration of mediator release. Adenosine did not appear to decrease the minimum concentration of ionophore required to stimulate release. The poorly metabolized adenosine analog 2-chloroadenosine (100 µM) potentiated release approximately 2-fold whereas dipyridamole, an inhibitor of mast cell uptake of adenosine, did not inhibit adenosine potentiation suggesting that adenosine actions on mast cells may be mediated by direct activation of cell-surface receptors. Adenosine modestly potentiated the fall in cyclic AMP levels occurring 30 to 60 sec after ionophore challenge, raising the possibility that potentiation may be mediated by alterations in cyclic AMP metabolism. Mast cells suspended in media containing 1 mM EGTA and no added calcium demonstrated adenosine potentiation of A23187-induced release indicating that the adenosine effect was not mediated by increased translocation of calcium into the cells. Adenosine potentiation was completely blocked by 100 µM theophylline, a known inhibitor of adenosine binding, with 50% inhibition of release occurring at 3 µM theophylline. These studies indicate that adenosine may be an important modulator of mast cell function in complex tissues and that part of the pharmacologic action of theophylline in human bronchial asthma may be through inhibition of adenosine binding, rather than direct effects on cyclic AMP metabolism.