It has been found that cyclic AMP and cyclic AMP-elevating agents inhibit formyl-methionyl-leucyl-phenylalanine (fMLP)-stimulated superoxide production from polymorphonuclear leukocytes (PMNs). The quantitative differences of this inhibitory effect on human and rabbit blood versus human salivary and rabbit peritoneal (tissue) PMNs were investigated. PMNs from all sources showed the same pattern of fMLP-stimulated superoxide generation, although it was slightly higher in tissue PMNs. However, treatment with salbutamol differentially blunted fMLP-stimulated superoxide production from blood PMNs compared with tissue PMNs in both human and rabbit. While it could inhibit production from blood PMNs by 30-60%, it had only a negligible effect on generation from tissue PMNs. Similarly, forskolin, phosphodiesterase IV inhibitor Re-201724, and dibutyryl cyclic AMP showed significantly higher inhibitory effects on superoxide generation from blood PMNs than tissue PMNs in both species. beta-Adrenergic receptors, cyclic AMP accumulation, and protein kinase A activity were investigated in blood versus tissue PMNs to clarify the mechanism underlying the above-mentioned differences. At the beta-adrenergic receptor level, no significant changes were detected in the number or the binding affinity of the receptors in tissue Versus blood PMNs of human and rabbit. On the other hand, cyclic AMP accumulation was significantly higher in response to salbutamol and Re-201724 in fMLP-stimulated blood versus tissue PMNs in human and rabbit. At the same time, blood PMNs showed significantly higher cyclic AMP-dependent protein kinase A activity than tissue PMNs in human and rabbit. We concluded that tissue PMNs are less responsive to the effect of cyclic AMP-elevating agents in terms of fMLP-stimulated superoxide inhibition. This is due to differences, at least, at two levels. The first is lower accumulation of cyclic AMP and the second is lower protein kinase A activity in tissue versus blood PMNs.