Several novel inhibitors of human synovial fluid phospholipase A2 (HSF-PLA2) were evaluated in cellular models of inflammatory mediator release (murine macrophage and human neutrophil) and topical in vivo inflammatory skin models in mice to ascertain the scope of effects which might be observed for PLA2 inhibitors. Potent inhibition of HSF-PLA2 in vitro can be observed with compounds such as scalaradial and ellagic acid, which both have IC50 values of 0.02 μM (using autoclaved [3H]-arachi-donic-acid (AA)-labelled Escherichia coli membranes as substrate). Luffariellolide, a manoalide analog, and aristolochic acid are less potent (IC50 = 5 and 46 μM, respectively) in this assay. An interesting observation is that ellagic acid in cellular assays does not inhibit macrophage eicosanoid production and only 30% inhibition of PAF biosynthesis can be obtained at 50 μM in the human neutrophil. Possibly due to its irreversible mechanism of action, scalaradial retained its potent activity in both the macrophage (IC50 for PGE2 production = 0.05 μM) and neutrophil assays (IC50 for PAF biosynthesis = 1 μM). Aristolochic acid is active in these cellular assays (macrophage IC50 = 2.5 μM and neutrophil IC50 = 100 μM), but is consistently less active than either scalaradial or luffariellolide. The relative potencies of these compounds were determined in several murine in vivo inflammatory models such as oxazolone contact hypersensitivity, AA-induced ear edema and phorbol ester (PMA)-induced ear edema. In the mouse model of oxazolone contact hypersensitivity, these PLA2 inhibitors have little effect ( ≤ 30% inhibition at 400 μg/ear) with scalaradial and luffariellolide being less effective than either aristolochic or ellagic acid. PMA-induced ear edema was effectively inhibited by scalaradial, luffariellolide and aristolochic acid (ED50 = 70, 50 and 50 μg/ear, respectively) whereas ellagic acid was less effective (ED50 = 230 μg/ear). In AA-induced ear edema, these PLA2 inhibitors had minimal effects, as would be expected for compounds which inhibit PLA2. These results, especially those of ellagic acid, suggest that caution should be taken in the extrapolation of potency against a purified human extracellular type PLA2 to the scope of activities these compounds might have in the cellular and in vivo models. The consistency of scalaradial and luffariellolide may be inherent to their irreversible mechanism of action, which is a factor to be accounted for in the extrapolation of enzyme data to cellular and in vivo models.