Abstract Neutrophilic polymorphonuclear leukocytes (PMNs) were incubated with opsonized zymosan and lysed human erythrocytes (RBCs) as measured by a 51Cr release method. Conversely, myeloperoxidase (MPO)-negative hydrogen peroxide (H2O2)-generating cells, derived in vitro from human monocytes (monocyte-derived cells (MDCs), were ineffective per se but capable of augmenting the lysis by PMNs. The lysis by PMNs and PMNs plus MDCs was inhibited by catalase, azide, taurine, and alanine, consistent with the requirement for hypochlorous acid (HOCl). As detected under conditions similar to those used for lytic assays, MDCs failed to produce HOCl but augmented the HOCl recovery from the PMN-RBC system. Moreover, when the extent of the lysis was plotted as a function of the HOCl recovery, a positive linear relationship was found. Although the actual size of the H2O2 extracellular pool could not be measured because of the inexistence of a reliable assay to probe our cytolytic model without perturbing the equilibrium of the system, the results presented suggest that MDCs enhance the PMN-mediated lysis by improving the HOCl production, presumably by supplying extra amounts of H2O2 to be handled by PMN MPO. In fact, the events mediated by MDCs could be reproduced by using an appropriate H2O2-generating enzymatic system (glucose-glucose oxidase). The present study provides direct evidence for the possibility of cooperation between MPO-positive and MPO-negative phagocytes in exerting functions (HOCl production and, in turn, cytolysis) possibly relevant to the outcome of inflammatory processes.