Endothelium produces oxygen-derived free radicals (nitric oxide, NO&z.rad;; superoxide anion, O(2)(*-)) which play a major role in physiology and pathology of the vessel wall. However, little is known about endothelium-derived O(2)(*-) production, particularly due to the difficulty in assessing O(2)(*-) when its production is low and to controversies recently raised about the use of lucigenin-enhanced chemiluminescence. We compared four techniques of O(2)(*-) assessment when its production is low. In the present study, we have compared ferricytochrome c reduction, electron spin resonance (ESR) spectroscopy using DMPO as spin trap, hydroethidine fluorescence, and lucigenin-enhanced chemiluminescence to assess O(2)(*-) production in cultured bovine aortic endothelial cells (BAEC). We focused our study on extracellular O(2)(*-) production because the specificity of the signal is provided by the use of superoxide dismutase, and this control cannot be obtained intracellularly. We found that the calcium ionophore A23187 dose-dependently stimulated O(2)(*-) production, with a good correlation between all four techniques. The signals evoked by postconfluent BAEC were increased 2- to 7-fold in comparison to just-confluent BAEC, according to the technique used. Ferricytochrome c 20 microm rather than at 100 microm appears more suitable to detect O(2)(*-). However, in the presence of electron donors such as NADH or NADPH, lucigenin-enhanced chemiluminescence generated high amounts of O(2)(*-). Thus, ferricytochrome c reduction, electron spin resonance (ESR), and hydroethidine fluorescence appear as adequate tools for the detection of extracellular endothelium-derived O(2)(*-) production, whereas lucigenin may be artifactual, even when a low concentration of lucigenin is employed.