To test the hypothesis that cardiac norepinephrine depletion related to heart failure alters contractile responses to beta-adrenergic agonists with a component of "indirect" action (acting by release of neuronal norepinephrine), we examined the inotropic potential of several pharmacologically distinct beta-agonists. Contractile responses to the nonselective beta-agonist isoproterenol, the beta 2-selective agonist zinterol, and the direct- and indirect-acting agonists dopamine and dopexamine were compared in isolated right ventricular trabeculae removed from failing, nonfailing innervated, and previously transplanted and, therefore, denervated nonfailing human hearts. In failing hearts, the contractile response to isoproterenol was significantly lower (41%) than that in nonfailing innervated hearts. The responses to the mixed agonists dopamine and dopexamine were even more attenuated in failing hearts, to a level 76-90% lower than those of nonfailing innervated hearts. In denervated, previously transplanted, nonfailing hearts, the contractile responses to the mixed agonists dopamine and dopexamine were 66-72% lower than those in the nonfailing innervated group, but the response to isoproterenol was not significantly different. The response to zinterol was not significantly different among the three groups. In subjects with severe heart failure, in vivo hemodynamic responses to dopexamine were compared with those of the direct-acting beta-agonist dobutamine. Responses to dopexamine and dobutamine were measured before and after prolonged continuous infusions of each drug. The response to dopexamine, but not to dobutamine, diminished over time. We conclude that a large component of the inotropic response to dopamine and dopexamine in human hearts is due to the ability of these agonists to promote the release of neuronal norepinephrine; when neuronal norepinephrine is depleted, indirect-acting agonists are less able to produce an inotropic response.