1. The effect on cyclic nucleotide contents of selective inhibitors of cyclic nucleotide phosphodiesterase (PDE) isoforms III and IV (respectively SK&F 94120 and rolipram) and their interactions with endothelium and NO have been studied in rat aorta in the presence of indomethacin (10 microM). The participation of NO was assessed by using either NG-nitro-L-arginine methyl ester (L-NAME) (NO synthase inhibitor: 30 microM) or 3-morpholinosydnonimine (SIN-1, NO donor: 10 microM with SOD 100 units ml-1). 2. The presence of endothelium significantly increased both adenosine 3':5'-cyclic monophosphate (cyclic AMP, 1.7 fold) and guanosine 3':5'-cyclic monophosphate (cyclic GMP, 2.2 fold) contents. Cyclic GMP was largely affected by L-NAME or SIN-1 treatment, this was not the case for cyclic AMP suggesting that the presence of endothelium modified cyclic AMP content in aorta independently of the NO production. 3. In the presence or absence of endothelium, neither SK&F 94120 nor rolipram, alone or combined, significantly modified cyclic GMP content. 4. The PDE III inhibitor significantly affected cyclic AMP content only in non treated aorta without endothelium. In contrast, the PDE IV inhibitor increased cyclic AMP in all conditions. These increases were generally about 2 fold but markedly higher in aorta treated with SIN-1 and superoxide dismutase (SOD, 6 fold). Association of a low concentration of the PDE III inhibitor (5 microM) with the PDE IV inhibitor (30 microM) potentiated the effect of the PDE IV inhibitor on cyclic AMP content, except for aorta without endothelium treated with SIN-1 plus SOD. 5. These data indicate that the presence of the endothelium could increase cyclic AMP content independently of NO and prostacyclin (PGI2) production. Furthermore, an increase in cyclic GMP content (modulated by NO production) could enhance the cyclic AMP accumulation induced by the PDE IV inhibitor. This result supports the hypothesis that PDE III inhibition by endogenous cyclic GMP may potentiate the effect of PDE IV inhibition on cyclic AMP content. Taken together with our previous studies on relaxation, these results suggest that the NO/cyclic GMP pathway could induce PDE IV-dependent regulation of cyclic AMP via PDE III inhibition.