Little is known about the role of peripheral blood mononuclear cells (PBMCs) in lipopolysaccharide (LPS) elimination. We studied the endotoxin elimination capacities (EEC) of PBMCs of 15 healthy volunteers, 13 patients with sepsis, and 1 patient suffering from paroxysmal nocturnal hemoglobinuria (PNH). Although expression of CD14, the best-characterized receptor for LPS to date, was reduced from 93.6% ± 0.8% in healthy subjects to 50.5% ± 6.5% in patients with sepsis and was 0.3% in a patient with septic PNH, EEC were found to be unchanged. There was no difference in the amount of tumor necrosis factor alpha (TNF-α) released by PBMCs of healthy donors and patients with sepsis. Anti-CD14 antibodies (MEM-18) completely suppressed EEC, binding of fluorescein isothiocyanate-labeled LPS to monocytes as determined by FACScan analysis, and TNF-α release in all three groups studied. The concentrations of soluble CD14 (sCD14) secreted by endotoxin-stimulated PBMCs from healthy donors and patients with sepsis amounted to 4.5 ± 0.4 and 20.1 ± 1.8 ng/ml, respectively. Based on our results, we suggest that PBMCs eliminate LPS by at least two different mechanisms; in healthy subjects, the membrane CD14 (mCD14) receptor is the most important factor for LPS elimination, while in patients with sepsis (including the septic state of PNH), increased sCD14 participates in LPS elimination. Secretion of sCD14 is strongly enhanced under conditions of low expression of mCD14 in order to counteract the reduction of mCD14 and maintain the function of monocytes. This sCD14 may substitute the role of mCD14 in LPS elimination during sepsis. The elimination of LPS by PBMCs correlates with the binding reaction and the secretion of TNF-α.

Assessment of circulating endotoxin during the perioperative period, which is only demonstrated by the Limulus amebocyte lysate (LAL) test, may be modulated by several endotoxin-binding proteins. Endotoxin-neutralizing capacity (ENC) and the plasma levels of soluble CD14 (sCD14), lipopolysaccharide-binding protein, and bactericidal/permeability-increasing protein (BPI) were determined in 40 patients 6 h prior to skin incision for major abdominal surgery. The bioactivity of plasma endotoxin was tested by the polymyxin B-inhibited stimulatory activity of the plasma samples on healthy monocytes as measured by the release of tumor necrosis factor alpha. Plasma endotoxin levels in almost all patients increased from 0.05 ± 0.01 to 0.23 ± 0.03 experimental units (EU) per ml (P < 0.001); more specifically, 17 of 40 samples showed endotoxin levels of greater than 0.2 EU per ml and corresponding reductions in ENC. Soluble CD14 plasma levels were decreased from 5.6 ± 0.3 to 4.6 ± 0.3 μg per ml (P < 0.05). ENC was strongly correlated with the sCD14 plasma concentration throughout the period of observation. The addition of sCD14-neutralizing monoclonal anti-sCD14 antibodies reduced ENC both pre- and postoperatively. No correlation could be established between ENC and the plasma levels of BPI, high-density lipoproteins, or low-density lipoproteins determined by measuring the concentrations of apoprotein A and apoprotein B. Biologically active endotoxin was found in only 6 of 17 samples with endotoxin levels greater than 0.2 EU per ml in the LAL test. These samples could be characterized by their perioperative loss of at least 35% of their sCD14. No change in sCD14 was detected in the remaining 11 samples. The perioperative loss of ENC is partly caused by the loss of sCD14 resulting from its consumption by endotoxin reaching the bloodstream. This study demonstrated the role of sCD14 on the bioactivity of circulating endotoxin in a human model of endotoxemia after major abdominal surgery.