Human neutrophils (PMN), when stimulated with such chemotaxins as phorbol myristate acetate (PMA), destroy erythrocytes and other targets. Cytotoxicity depends on PMN-generated reactive oxygen metabolites, yet the exact toxic specie and its mode of production is a matter of some dispute. Using 51Cr-labeled erythrocytes as targets, we compared various reactive-O2 generating systems for their abilities to lyse erythrocytes as well as to oxidize hemoglobin to methemoglobin. PMA-activated PMNs or xanthine oxidase plus acetaldehyde were added to target erythrocytes in amounts that provided similar levels of superoxide. PMNs lysed 68.3±2.9% (SEM) of targets, whereas the xanthine oxidase system was virtually impotent (2.3±0.8%). In contrast, methemoglobin formation by xanthine oxidase plus acetaldehyde was significantly greater than that caused by stimulated PMNs (P<0.001). A similar dichotomy was noted with added reagent H2O2 or the H2O2-generating system, glucose plus glucose oxidase; neither of these caused 51Cr release, but induced 10-70% methemoglobin formation. Thus, although O2̄ and H2O2 can cross the erythrocyte membrane and rapidly oxidize hemoglobin, they do so evidently without damaging the cell membrane. That a granule constituent of PMNs is required to promote target cell lysis was suggested by the fact that agranular PMN cytoplasts (neutroplasts), although added to generate equal amounts of O2̄ as intact PMNs, were significantly less lytic to target erythrocytes (P<0.01). Iron was shown to be directly involved in lytic efficiency by supplementation studies with 2 μM iron citrate; such supplementation increased PMN cytotoxicity by ~ 30%, but had much less effect on erythrocyte lysis by neutroplasts (~3% increase), and no effect on lysis in the enzymatic oxygen radical-generating systems. These results suggest a critical role for an iron-liganding moiety that is abundantly present in PMN, marginally so in neutroplasts, and not at all in purified enzymatic systems - a moiety that we presume catalyzes very toxic O2 specie generation in the vicinity of juxtaposed erythrocyte targets. The obvious candidate is lactoferrin (LF), and indeed, antilactoferrin IgG, but not nonspecific IgG, reduced PMN cytotoxicity by >85%. Re-adding 10-8 M pure LF to neutroplasts increased their ability to promote hemolysis by 48.4±0.9% - to a level near that of intact PMNs. We conclude that O2̄ and H2O2 are not sufficient to mediate target cell lysis, but require iron bound to LF, which, in turn, probably generates and focuses toxic O2 radicals, such as OH, to target membrane sites.