Role for macrophage products in endotoxin-induced polymorphonuclear leukocyte accumulation during inflammation

Abstract

Endotoxins (lipopolysaccharide, LPS) released by Gram-negative bacteria induce acute inflammation with polymorphonuclear leukocyte (PMNL) infiltration. The mechanism of PMNL accumulation appears to be complement-independent and is not well understood. Here, we report investigation of the factors which may mediate LPS-induced PMNL accumulation in the pleural cavity and skin of rabbits. The intrapleural injection of 50 ng of Escherichia coli 0111 LPS caused the appearance in the exudate fluid of an activity which, upon intradermal injection induced PMNL accumulation in the skin, as measured by a 51Cr-labeled leukocyte assay and which was confirmed histologically. This activity preceded by 30 minutes the massive influx of PMNL into the pleural cavity. 125I-labeled LPS, gel filtration chromatography, limulus amebocyte lysate assays, and polymyxin B allowed distinction between reactions in the skin attributable to LPS and reactions due to the effect of this "PMNL infiltration-inducing activity." Pleural macrophages cultured for 3 to 6 hours with 3 to 30 ng/ml of LPS also released factors which induced PMNL infiltration into the skin. Sephadex G-100 chromatography of LPS-induced pleural exudate fluid or of supernatants from LPS-stimulated macrophage cultures yielded identical elution profiles, with one major peak of PMNL infiltration-inducing activity at an apparent molecular weight of 45,000 and a minor peak at 14,000 to 18,000. Only the low molecular weight fraction contained interleukin 1 activity. Lipid A was required for the secretion of these factors by macrophages. The LPS shed by killed E. coli also induced macrophage production of PMNL infiltration-inducing activity. The activity was sensitive to pronase, and its production was inhibited by an inhibitor of protein synthesis (cycloheximide). The active factors did not induce PMNL chemotaxis, aggregation, or chemiluminescence in vitro indicating that the activity was not C5a. We conclude that PMNL infiltration induced by LPS and perhaps by Gram-negative bacteria, may be mediated in part by the secretion from tissue macrophages of factors which can recruit PMNLs from the blood. The most active of these (approximately equal to 45,000 daltons) lacks interleukin-1 or PMNL chemotactic activity.