The cyclic AMP phosphodiesterase (cAMP PDE) inhibitor and cardiotonic agent lixazinone (N-cyclohexyl-N-methyl-4-[(1,2,3,5-tetrahydro-2- oxoimidazo[2,1-b]quinazolin-7-yl)oxy]butyramide, RS-82856, 1) and its acid and base addition salts were found to be insufficiently soluble in formulations suitable for intravenous administration. These results prompted an investigation into potential prodrugs with enhanced aqueous solubility designed to deliver 1 by three distinct mechanisms: (1) decarboxylation of alpha-carboxamides; (2) hydrolytic loss of a solubilizing N-1-(acyloxy)methyl or (N,N-dialkylamino)methyl moiety; or (3) intramolecular closure of a guanidino ester or amide. The target compounds were evaluated as delivery systems for 1 by three criteria: (1) chemical conversion rate to 1 under physiological conditions; (2) inhibition of type IV cAMP PDE at a fixed time point; and (3) in vivo inotropic activity in anesthetized dogs by both intravenous and oral administration. Release of 1 from 4a (series 1) was found to be too slow to be of value as a prodrug of 1, since decarboxylation could be induced only by strong acid, conditions under which hydrolytic ring opening was found to severely compete. Conversely, 1 was released too readily on exposure of (N,N-dialkylamino)methyl derivatives such as 8d (series 2) to physiological conditions, although no large increase in aqueous solubility was realized. Finally, both the physicochemical and in vitro studies indicated that ring closure of the guanidinium esters and amides 17a-k (series 3) to 1 was quantitative and pH- and time-dependent, suggesting the possibility of delivery of the open, water-soluble prodrug form, followed by closure to 1 in plasma. Detailed examination of these agents in vivo, however, demonstrated that only those compounds that rapidly cyclized to 1, as measured by plasma levels of 1, exhibited inotropic activity, indicating that the open prodrug form was not efficiently absorbed upon oral administration.