1. During in vitro culture in 10% human AB serum, human peripheral blood monocytes acquire a macrophage-like phenotype. The underlying differentiation was characterized by increased activities of the macrophage marker enzymes unspecific esterase (NaF-insensitive form) and acid phosphatase, as well as by a down-regulation in surface CD14 expression. 2. In parallel, a dramatic change in the phosphodiesterase (PDE) profile became evident within a few days that strongly resembled that previously described for human alveolar macrophages. Whereas PDE1 and PDE3 activities were augmented, PDE4 activity, which represented the major cyclic AMP-hydrolysing activity of peripheral blood monocytes, rapidly declined. 3. Monocytes and monocyte-derived macrophages responded to lipopolysaccharide (LPS) with the release of tumour necrosis factor-alpha (TNF). In line with the change in CD14 expression, the EC50 value of LPS for induction of TNF release increased from approximately 0.1 ng ml-1 in peripheral blood monocytes to about 2 ng ml-1 in macrophages. 4. Both populations of cells were equally susceptible towards inhibition of TNF release by cyclic AMP elevating agents such as dibutyryl cyclic AMP, prostaglandin E2 (PGE2) or forskolin, which all led to a complete abrogation of TNF production in a concentration-dependent manner and which were more efficient than the glucocorticoid dexamethasone. 5. In monocytes, PDE4 selective inhibitors (rolipram, RP73401) suppressed TNF formation by 80%, whereas motapizone, a PDE3 selective compound, exerted a comparatively weak effect (10-15% inhibition). Combined use of PDE3 plus PDE4 inhibitors resulted in an additive effect and fully abrogated LPS-induced TNF release as did the mixed PDE3/4 inhibitor tolafentrine. 6. In monocyte-derived macrophages, neither PDE3- nor PDE4-selective drugs markedly affected TNF generation when used alone (< 15% inhibition), whereas in combination, they led to a maximal inhibition of TNF formation by about 40-50%. However, in the presence of PGE2 (10 nM), motapizone and rolipram or RP73401 were equally effective and blocked TNF release by 40%. Tolafentrine or motapizone in the presence of either PDE4 inhibitor, completely abrogated TNF formation in the presence of PGE2. Thus, an additional cyclic AMP trigger is necessary for PDE inhibitors to become effective in macrophages. 7. Finally, the putative regulatory role for PDE1 in the regulation of TNF production in macrophages was investigated. Zaprinast, at a concentration showing 80% inhibition of PDE1 activity (100 micromol l-1), did not influence TNF release. At higher concentrations (1 mmol l-1), zaprinast became effective, but this inhibition of TNF release can be attributed to a significant inhibitory action of this drug on PDE3 and PDE4 isoenzymes. 8. In summary, the in vitro differentiation of human peripheral blood monocytes to macrophages is characterized by a profound change in the PDE isoenzyme pattern. The change in the PDE4 to PDE3 ratio is functionally reflected by an altered susceptibility towards selective PDE inhibitors under appropriate stimulating conditions.