We tested the hypothesis that the negative inotropic effect (NIE) of cocaine is mediated, at least in part, by cholinergic stimulation and can be correlated with the degree of adenosine 3',5'-cyclic monophosphate (cAMP) dependency of the inotropic state. Cardiac myocytes were isolated from left ventricles of ferrets and loaded with the fluorescent Ca2+ indicator indo 1. Cells were placed in physiological solution containing 2.0 mM Ca2+ and stimulated at 0.5 Hz and 30 degrees C. Cocaine decreased peak cell shortening and peak intracellular Ca2+ in a concentration-dependent manner (10(-8)-10(-4) M). The concentration-response curve of cocaine was shifted significantly downward compared with those of lidocaine and procaine in the same range of concentrations. Atropine (10(-6) M) shifted the concentration-response curve of cocaine, but not those of lidocaine and procaine, rightward, with a pA2 value (7.66) similar to that obtained with carbachol (7.99). With prior addition of isoproterenol (ISO, 10(-8) M) or increased Ca2+ (4.0 mM) to increase cell shortening to the same degree (approximately 60%), cocaine and carbachol decreased contractility to a significantly greater extent in ISO-stimulated myocytes. To clarify whether these treatments changed responsiveness of the contractile elements to Ca2+, the effect of 2,3-butanedione monoxime, an agent that interferes with the interaction of myosin and actin, was tested with previous addition of ISO or increased Ca2+, and no differential effect occurred. Therefore, we postulate that 1) the NIE of cocaine on myocytes is caused by decreased Ca2+ availability; 2) this effect is due to specific stimulation of cholinergic receptors in addition to other direct myocardial (probably local anesthetic) effects; and 3) the NIE correlates with the level of cAMP dependence of the inotropic state.