Michael A. Fisher

Activation of Kupffer cells and neutrophils for reactive oxygen formation is responsible for endotoxin-enhanced liver injury after hepatic ischemia.

The potential role of reactive oxygen species generated by Kupffer cells and neutrophils was investigated in a model of endotoxin-enhanced liver injury after hepatic ischemia. Male Fischer rats were subjected to 20 min ischemia and reperfusion of up to 24 h; .5 mg/kg Salmonella enteritidis endotoxin was injected at 30 min of reperfusion. The animals developed severe liver injury resulting in 50% hepatocellular necrosis at 24 h. Isolated Kupffer cells and neutrophils from the postischemic liver generated 10-fold more superoxide than cells from control livers. Treatment with gadolinium chloride (GdCl3) selectively reduced the capacity of Kupffer cells to generate superoxide by 65% and attenuated liver injury by 73% at 4 h and 58–69% at 24 h. Monoclonal antibodies against neutrophil adhesion molecules (CD11/CD18) had no effect on the early injury but reduced hepatocellular necrosis by 90–95% at 24 h. The antioxidant Trolox and the iron-chelator deferoxamine attenuated liver injury by 71 and 80%, respectively. It is concluded that Kupffer cells are mainly responsible for the initial injury, and neutrophils are the dominant cytotoxic cell type during the later phase. Reactive oxygen generated by both cell types is critical for this pathogenesis.

Lipid peroxidation as molecular mechanism of liver cell injury during reperfusion after ischemia.

The pathophysiological importance of reactive oxygen species has been extensively documented in the pathogenesis of hepatic ischemia-reperfusion injury. Kupffer cells and neutrophils were identified as the dominant sources of the postischemic oxidant stress. To test the hypothesis that a direct free radical-mediated injury mechanism (lipid peroxidation; LPO) may be involved in the pathogenesis, highly sensitive and specific parameters of LPO, i.e., hydroxy-eicosatetraenoic acids (HETES), and F2-isoprostanes, were determined by gas chromatographic-mass spectrometric analysis in liver tissue and plasma during 45 min of hepatic ischemia and up to 24 h of reperfusion. A significant 60-250% increase of F2-isoprostane levels in plasma was found at all times during reperfusion; the HETE content increased only significantly at 1 h of reperfusion and in severely necrotic liver tissue at 24 h with increases between 90-320%. On the other hand, in a model of LPO-induced liver injury (infusion of 0.8 mumol tert-butylhydroperoxide/min/g liver), the hepatic HETE content increased two to fourfold over baseline values at 45 min, i.e., before liver injury. A further increase to 12- to 30-fold of baseline was observed during moderate liver injury. Based on these quantitative comparisons of LPO and liver injury, it seems highly unlikely that LPO is the primary mechanism of parenchymal cell injury during reperfusion, although it cannot be excluded that LPO may be important as a damaging mechanism in a limited compartment of the liver, e.g., endothelial cells, close to the sources of reactive oxygen, e.g., Kupffer cells and neutrophils.

Sequestration Of Neutrophils In The Hepatic Vasculature During Endotoxemia Is Independent Of β2 Integrins And Intercellular Adhesion Molecule-1

Antibodies against cellular adhesion molecules protect against neutrophil-induced hepatic injury during ischemia-reperfusion and endotoxemia. To test if p2 integrins on neutrophils and intercellular adhesion molecule-1 (ICAM-1) on endothelial cells are involved in neutrophil sequestration in the hepatic vasculature, neutrophil accumulation in the liver was characterized during the early phase of endotoxemia. Intravenous injection of Salmonella enteritidis endotoxin induced a dose-dependent activation of complement, tumor necrosis factor-α (TNF-α) formation, and an increase of hepatic neutrophils with maximal numbers at 5 mg/kg (90 min: 339 ± 14 cells/50 high power fields; controls: 18 ± 2). Administration of 15 μg/kg TNF-α and intravascular complement activation with cobra venom factor (75 αg/kg) had additive effects on hepatic neutrophil accumulation compared with each mediator alone. Monoclonal antibodies (2 mg/kg) to ICAM-1 and the a-chain (CD11a, CD11 b) or the β-chain (CD18) of β2 integrins had no significant effect on hepatic neutrophil count after endotoxin. In contrast, these antibodies inhibited peritoneal neutrophil infiltration in response to glycogen administration by 28% (CD11b), 60% (CD11a, ICAM-1), and 92% (CD18). Our data suggest that TNF-α and complement factors contribute to hepatic neutrophil sequestration during the early phase of endotoxemia. Despite the fact that these inflammatory mediators can up-regulate integrins and ICAM-1, these adhesion molecules are not necessary for neutrophil accumulation in hepatic sinusoids. The protective effect of these antibodies against neutrophil-induced liver injury appears to be due to inhibition of transendothelial migration and adherence to parenchymal cells.

Inhibition of NF-kB activation by dimethyl sulfoxide correlates with suppression of TNF-α formation, reduced ICAM-1 gene transcription and protection against endotoxin-induced liver injury.

The effect of the free radical scavenger dimethyl sulfoxide (DMSO) on activation of the nuclear transcription factor KB (NF-KB) was investigated in an experimental model of endotoxin-induced liver failure. In galactosamine-sensitized C3Heb/FeJ mice, DMSO (10 mL/kg) effectively inhibited endotoxin-induced hepatic NF-KB activation, suppressed TNF-α levels in plasma by 86%, attenuated intercellular adhesion molecule-1 (ICAM-1) mRNA formation, blocked hepatic neutrophil accumulation by 79%, and reduced liver injury by 80%. In galactosamine-sensitized mice treated with 20 /μg/kg murine TNF-α, DMSO moderately reduced hepatic NF-KB and decreased ICAM-1 mRNA formation and liver injury by 83%, but had no significant effect on hepatic neutrophil accumulation. Thus, DMSO was able to inhibit, at least in part, two critical NF-KB-dependent steps in the pathophysiology, i.e., TNF-α formation and ICAM-1 gene transcription. Our data suggest the involvement of redox-sensitive events in the signal transduction pathway of NF-KB activation in the liver. Inhibition of NF-KB activation correlates with the reduced activation of proinflammatory genes in vivo and the subsequent attenuation of inflammatory liver injury. Thus, antioxidants that are NF-KB inhibitors may have therapeutic potential in endotoxin shock and sepsis.

Increased P‐selectin gene expression in the liver vasculature and its role in the pathophysiology of neutrophil‐induced liver injury in murine endotoxin shock

We studied the role of P‐selectin, an adhesion molecule known to be important for neutrophil localization to sites of inflammation, in a model of inflammatory liver injury. Male C3Heb/ FeJ (ET‐sensitive) mice treated with 700 mg/kg galactosamine and 100 μg/kg Salmonella abortus equi endotoxin (Gal/ET), murine tumor necrosis factor α (TNF‐α,15 μg/kg), or interleukin‐1 (IL‐1, 13–23 μg/kg), showed increased P‐selectin mRNA levels in the liver. In contrast, C3H/HeJ (ET‐resistant) mice responded only to cytokines with P‐selectin mRNA formation. Whereas no P‐selectin expression was detectable in control livers, there was temporary staining of endothelium in large blood vessels but not in sinusoids between 3 and 5 h after ET, TNF‐α, or IL‐1 treatment. Severe liver injury induced by Gal/ET at 7 h was not inhibited by an anti‐P‐selectin antibody in C3Heb/FeJ mice or in P‐selectin‐deficient animals. Sequestration of neutrophils in sinusoids, i.e. those neutrophils that have been identified as critical for the injury, was not affected by the antibody or in P‐selectin‐deficient mice. However, the temporary margination in portal and postsinusoidal venules was reduced by 75% in anti‐P‐selectin antibody‐treated animals and by 51% in P‐selectin‐deficient mice. We conclude that hepatic P‐selectin gene transcription in vivo involves cytokines. However, blocking P‐selectin neither attenuated sinusoidal neutrophil sequestration nor prevented neutrophil‐induced liver injury during endotoxin shock but attenuated neutrophil margination in larger vessels. Thus, our data demonstrate similarities and fundamental differences in the requirement for adhesion molecules to localize neutrophils in the liver vasculature compared to other organs during an inflammatory response. J. Leukoc. Biol. 63: 288–296; 1998.

Sinusoidal endothelial cell and parenchymal cell injury during endotoxemia and hepatic ischemia-reperfusion: protection by the 21-aminosteroid tirilazad mesylate

Abstract The early vascular injury in the liver was characterized in an experimental model of multiple organ failure (MOF). Significant increases of hyaluronic acid levels (660%) and plasma alanine aminotransferase activities (1050%) were observed after 20 min hepatic ischemia followed by 4 h reperfusion and injection of 0.5 mg/kg Salmonella enteritidis endotoxin at 30 min reperfusion. Morphological evaluation of sinusoids with transmission electron microscopy indicated neutrophil and Kupffer cell activation as well as damage or loss of sinusoidal endothelial cells. Hepatocellular injury was evident from fused microvilli and blebbed plasma membranes. Treatment with the 21-aminosteroid tirilazad mesylate (U-74006F) (2 × 3 mg/kg) reduced plasma hyaluronic acid levels by 61% and plasma transaminase activities by 69% suggesting a beneficial effect on sinusoidal endothelial cell and parenchymal cell injury. This was confirmed by morphology. Our data provide morphological and functional evidence for severe injury to sinusoidal endothelium and the vascular pole of hepatocytes in this model of MOF. U-74006F significantly protected the liver against this Kupffer cell- and neutrophil-mediated injury. Thus, U-74006F may be a promising therapeutic for liver dysfunction and failure during a local or systemic inflammatory response.

Priming of phagocytes for reactive oxygen production during hepatic ischemia-reperfusion potentiates the susceptibility for endotoxin-induced liver injury

Plasma levels of glutathione disulfide (GSSG) as an indicator of a vascular oxidant stress, tumor necrosis factor-alpha (TNF-alpha) formation, and liver injury (alanine aminotransferase activity, histology) were monitored in male Fischer rats after 30 min of hepatic ischemia followed by up to 4 hr of reperfusion. The injection of 1 mg/kg Salmonella enteritidis endotoxin at 30 min of reflow potentiated the postischemic oxidant stress and liver injury. TNF-alpha levels increased from 10 +/- 7 pg/ml (baseline) to 3,553 +/- 738 pg/ml after ischemia-reperfusion followed by endotoxin, or to 3,670 +/- 508 pg/ml after endotoxin alone. Depletion of serum complement before ischemia attenuated the endotoxin-mediated increase of reactive oxygen formation by 70% but did not affect TNF-alpha levels. Complement activation with cobra venom factor (CVF) during reperfusion had an effect similar to that of endotoxin on the oxidant stress and liver injury. CVF did not increase TNF-alpha formation during reperfusion. Kupffer cells and neutrophils isolated from the postischemic liver 2.5 hr after endotoxin injection generated 600% and 400% more superoxide, respectively, than cells isolated from control livers. The results demonstrate a substantial priming of hepatic phagocytes for reactive oxygen production but not TNF-alpha formation, even after short periods of hepatic ischemia, and the vulnerability of the postischemic liver to severe endotoxin-induced injury. Activated complement seems to be mainly responsible for the effects. These results may explain the high risk for hepatic failure after extensive liver resection and hypovolemic shock.