The typical reverse passive Arthus reaction (RPA) was attained in rats by the instillation of a rabbit antiovalbumin serum into the lungs and intravenous injection of ovalbumin. Instillation of antiserum alone caused accumulation of polymorphonuclear leukocytes (PMN) and increased vascular permeability, but did not cause hemorrhage. However, when an intravenous injection of ovalbumin was also given, the vascular permeability of the lungs increased dramatically and PMN, as well as hemoglobin, were measurable in the lung lavage fluids by 4 hr after initiation of the reaction. Various proteinase inhibitors were instilled into the lungs after the initial stages of the RPA had developed, specifically to investigate their effect on the development of the hemorrhage, which we chose to monitor as an indicator of severe vascular damage. A cephalosporin-based beta-lactam, L-658,758, which is a time-dependent inhibitor of human and rat PMN elastase, effectively prevented the lung hemorrhage associated with the RPA reaction (ED50 = 2 x 55 micrograms doses/animal when instilled at 1.5 and 2.5 hr after initiating the RPA). The PMN elastase inhibitor, methoxysuccinyl-alanyl-alanyl-prolyl-valine-chloromethylketone, also inhibited hemorrhage in this model. Compounds of the same chemical class as these elastase inhibitors, but having no activity against PMN elastase in vitro, did not affect the hemorrhage associated with the RPA. Several specific inhibitors of proteinases other than PMN elastase (e.g., pepstatin and methoxysuccinyl-prolyl-glycyl-alanyl-lysine-chloromethylketone) were found to have little effect on the hemorrhage associated with the RPA reaction.(ABSTRACT TRUNCATED AT 250 WORDS)