Tumour necrosis factor α and interleukin 1β mediate lipopolysaccharide-induced endothelial monolayer permeability in whole blood

Abstract Background Overproduction of cytokines may be responsible for irreversible and systemic vascular damage observed among patients suffering from the sepsis syndrome. The resultant excessive fluid loss from the circulation may cause tissue hypoperfusion, organ failure and death. A whole blood assay and an in vitro endothelial model have been used to simulate permeability changes during the acute phase of Gram-negative infection. Methods Freshly drawn heparinized human blood from healthy volunteers was treated with bacterial lipopolysaccharide (LPS; Escherichia coli B55: 05). Plasma was then collected and incubated on human umbilical vein endothelial cell (HUVEC) monolayers cultured on semipermeable Transwell COL membranes. The permeability of the monolayers was determined using fluorescein isothiocyanate (FITC) albumin and rhodamine-β-isothiocyanate (RITC) dextran. The production of tumour necrosis factor (TNF) α and interleukin (IL) 1β in whole blood was measured using radioimmunossay. Results Incubation of whole blood with as little as 10 pg ml−1 LPS was sufficient to yield plasma that significantly increased endothelial permeability. This effect was dose dependent up to 10 ng ml−1 LPS, at which the permeability exceeded about seven times the permeability of control monolayers. In contrast, significant stimulation of endothelial permeability by LPS alone (i.e. without prior incubation in blood) required more than 50 ng ml−1 LPS. The effect of LPS pretreatment on endothelial permeability was also time dependent: at least 2–4 h of LPS preincubation was required, while maximum permeability was reached at longer preincubations (8–24 h). At these amounts and incubation times LPS was also capable of inducing TNF-α and IL-1β in whole blood. When antibodies directed against TNF-α and IL-1β were added to LPS-conditioned plasma before incubation on HUVEC monolayers the permeability increase was partly inhibited. However, almost complete inhibition was achieved when anti-TNF-α and anti-IL-1β were both added to LPS-conditioned plasma. Similar effects of LPS-conditioned plasma and anti-TNF-α and anti-IL-1β treatment were observed on the expression of E-selectin and intracellular adhesion molecule 1 by endothelial cells. Conclusion Whole blood potentiates the inflammatory response of endothelial cells to physiologically relevant doses of LPS. This potentiation is achieved by the production of both TNF-α and IL-1β. If this in vitro endothelial model represents the events observed in patients suffering from Gram-negative sepsis, these data indicate that protection requires simultaneous administration of inhibitors to TNF-α and IL-1β.