Development of endotoxin tolerance in humans in vivo.

OBJECTIVE: Endotoxin (lipopolysaccharide [LPS]) tolerance is characterized by a reduced responsiveness to a subsequent LPS challenge. In animal and human in vitro experiments, LPS tolerance is associated with an attenuated response of proinflammatory cytokines and an enhanced production of anti-inflammatory cytokines. It is unclear if this mechanism accounts for the development of LPS tolerance in humans in vivo. DESIGN: Clinical experimental study. SETTING: Intensive care research unit. PATIENTS: Fourteen healthy male volunteers. INTERVENTIONS: Intravenous injections of 2 ng/kg/day Escherichia coli LPS on 5 consecutive days. MEASUREMENTS AND MAIN RESULTS: Symptom scores, vital signs, leukocyte (elastase) and endothelial cell activation (von Willebrand factor [vWF]), and circulating cytokine levels. On day 1, the symptom score increased to 6.1 +/- 3.1, temperature to 37.8 +/- 0.4 degrees C, heart rate to 103 +/- 6/min (p < 0.0001 for all parameters) compared with 0.3 +/- 0.6, 36.2 +/- 0.5 degrees C, 79 +/- 4/min on day 5, respectively (p < 0.0001 between days 1 and 5). On day 1, elastase, vWF, and all cytokine levels increased significantly (p < 0.001 for all, except transforming growth factor (TGF)-beta, p = 0.02), whereas on day 5, this increase was significantly attenuated (p < 0.001) for elastase (61% +/- 6%), vWF (68% +/- 5%), tumor necrosis factor (97% +/- 3%), interleukin (IL)-6 (88% +/- 8%), IL-10 (87% +/- 7%), and IL-1ra (93% +/- 9% p = 0.018) but not for TGF-beta (5% +/- 22% p = 0.22). The tumor necrosis factor-alpha/IL-10 ratio showed an initial proinflammatory phase, followed by an anti-inflammatory phase on the first day. The proinflammatory phase was attenuated with 95% +/- 2%, whereas the reduction of the anti-inflammatory phase, without TGF-beta levels, was 99% +/- 1% on day 5 (p = 0.13 between phases). CONCLUSIONS: Endotoxin tolerance developed during five consecutive LPS administrations as demonstrated by the attenuated release of proinflammatory cytokines on the fifth day and was associated with less leukocyte and endothelial activation. In contrast to animal and human in vitro data, the attenuated response was not limited to the proinflammatory response, as a similar reduction in the anti-inflammatory cytokines was observed, with the exception of TGF-beta.