We previously demonstrated that right atrial strips from patients treated with Pi-selective antagonists exhibit sensitization of beta(1)-adrenergic responses in vitro. We also showed that cardiac beta(2)-adrenergic sensitization can be induced in normal subjects prospectively by beta(1)-blocker treatment. To determine whether such cross-talk could be induced in vitro, we studied beta-adrenoceptor-mediated vasorelaxation in deendothelialized rings of human coronary artery from patients undergoing cardiac transplantation. After incubation with 10 mu M phenoxybenzamine for 1 h, concentration-effect curves were determined to norepinephrine (NE) and epinephrine (EPI), with or without 300 nM CGP 20712A (a beta(1)-selective antagonist), 50 nM ICI 118551 (a beta(2)-selective antagonist), or both antagonists. Both beta(1)- and beta(2)-adrenergic components to vasorelaxation were detected. Other rings were incubated for 16 h with either 1 mu M NE (a selective beta(1)-adrenoceptor agonist) or 300 nM CGP 20712A, or both. After washout, concentration-effect curves were determined to EPI in the presence of 300 nM CGP 20712A (beta(2)-adrenergic responses). No differences in beta(2)-adrenergic vasorelaxation were noted after prolonged incubation with either CGP 20712A or the combination of CGP 20712A and NE. However, after incubation with 1 mu M NE, beta(2)-adrenergic vasorelaxation was desensitized, with a threefold reduction in potency (p < 0.05) and a 45% decrease in efficacy (p < 0.05); this occurred with no change in beta(1)-adrenergic vasorelaxation, as assessed by concentration-effect curves to NE in the presence of 50 nM ICI 118551. These results provide the first evidence that cross-talk between beta(1)- and beta(2)-adrenoceptors can be induced in vitro. The cross-desensitization observed may be the inverse effect of the previously observed beta(2)-adrenergic sensitization in response to chronic beta(1)-blockade.