Interleukin-6 infusion blunts proinflarnmatory cytokine production without causing systematic toxicity in a swine model of uncontrolled hemorrhagic shock

Background. Serum elevations of interleukin-6 (IL-6) correlate with multiple organ dysfunction syndrome and mortality in critically injured trauma patients. Data from rodent models of controlled hemorrhage suggest that recombinant IL-6 (rIL-6) infusion protects tissue at risk for ischemia-reperfusion injury. Exogenous rIL-6 administered during shock appears to abrogate inflammation, providing a protective rather than a deleterious influence. In an examination of this paradox, the current study aimed to determine whether rIL-6 decreases inflammation in a clinically relevant large animal model of uncontrolled hemorrhagic shock, (UHS), and to investigate the mechanism of protection. Methods: Swine were randomized to four groups (8 animals in each): (1) sacrifice, (2) sham (splenectomy followed by hemodilution and cooling to 33degreesC), (3) rIL-6 infusion (sham plus UHS using grade 5 liver injury with packing and resuscitation plus blinded infusion of rIL-6 [10 mcg/kg]), and (4) placebo (UHS plus blinded vehicle). After 4 hours, blood was sampled, estimated blood loss determined, animals sacrificed, and lung harvested for RNA isolation. Quantitative reverse transcriptase-polymerase chain reaction was used to assess granulocyte colony-stimulating factor (G-CSF), IL-6, and tumor necrosis factor-alpha (TNFalpha) messenger ribonucleic acid (mRNA) levels. Serum levels of IL-6 and TNFa were measured by enzyme-linked immunoassay (ELISA). Results: As compared with placebo, IL-6 infusion in UHS did not increase estimated blood loss or white blood cell counts, nor decrease hematocrit or plate-let levels. As compared with the sham condition, lung G-CSF mRNA production in UHS plus placebo increased eightfold (*p < 0.05). In contrast, rIL-6 infusion plus UHS blunted G-CSF mRNA levels, which were not significantly higher than sham levels (p = 0.1). Infusion of rIL-6 did not significantly affect endogenous production of either lung IL-6 or mRNA. As determined by ELISA, rIL-6 infusion did not increase final serum levels of IL-6 or TNFa over those of sham and placebo conditions. Conclusions: Exogenous rIL-6 blunts lung mRNA levels of the proinflammatory cytokine G-CSF. The administration of rIL-6 does not increase the local expression of IL-6 nor TNFa mRNA in the lung. Additionally, rIL-6 infusion does not appear to cause systemic toxicity.