Soluble phospholipase A2 has been implicated in the pathogenesis of local and systemic inflammatory reactions. Elevated levels of circulating phospholipase A2 (PLA2) correlate with the severity of circulatory collapse and pulmonary dysfunction in gram-negative septic shock. Characterization of septic shock serum PLA2 revealed a calcium-dependent enzyme with absolute 2-acyl specificity with a pH optimum of 7.5. We tested a number of therapeutic agents for their ability to inhibit PLA2 from human septic shock serum. Chloroquine, chlorpromazine, dexamethasone base, dexamethasone sodium phosphate, indomethacin, lidocaine, oleic acid, palmitic acid, promethazine, trans-retinoic acid, rutin and dl-alpha-tocopherol were all studied over the range of 10(-2) to 10(-7) M. All agents, with the sole exception of dexamethasone base, inhibited PLA2 activity at concentrations greater than 10(-3) M. PLA2 inhibition by dexamethasone sodium phosphate was factitious, due to the formation of calcium-phosphate complexes. Of the 11 agents studied, chlorpromazine was the most effective, with an IC50 of 7.5 X 10(-5) M, a membrane concentration achievable within its therapeutic range. Inhibition was non-competitive with an apparent Ki of 5 nM. Since serum PLA2 levels correlate with mortality in both experimental endotoxemia and clinical gram-negative septic shock, and chlorpromazine was previously shown to improve survival in these conditions, we postulate that its therapeutic efficacy resides at least in part in its PLA2-inhibitory activity.