Hepatology highlights


  • Neil Kaplowitz Liver Biology and Pathobiology Editor

  • Potential conflict of interest: Nothing to report.

Sorting Junk Is Useful

The methionine-choline deficient (MCD) diet model is widely used to study the pathogenesis of NASH. JNK has been implicated in development of obesity and insulin resistance. Two types of JNK (1+2) are expressed in liver. With the availability of JNK1 and JNK2 null mice, Schattenberg et al. examined the role of JNK in 4 week MCD diet induced steatohepatitis. MCD diet caused sustained increased hepatic JNK activity, phospho-JNK phospho-c-jun, total c-jun, and AP-1 DNA binding in wild type mice. Using knockout mice, it was shown that these effects were mainly due to JNK1. JNK1, but not JNK2, null mice developed markedly less steatohepatitis (see Fig.) accompanied by less triglyceride, pro-inflammatory gene expression, serum ALT, and apoptosis along with an increase in serum adiponectin but no difference in TNF. This is an important study which addresses the relative role of two forms of JNK. In contrast to these finding, recently presented work has suggested that JNK2 plays a more important role in TNF/galactosamine and acetaminophen toxicity. Thus, the specific targets of JNK1 and 2, the cell types involved, and the triggers for their activation in various liver disease models is an important subject for further exploration. (See HEPATOLOGY 2006;43:163–172.)

Illustration 1.

Another Weapon In the Spread of Death

Nuclear DNA breakdown by deoxyribonuclease 1 (DNASE1) is a common feature of necrotic cell death but is this important in the evolution toward cell death or a debris management process? Using the acetaminophen (APAP) hepatotoxicity model, Napirei et al. address this issue in wild type versus Dnasel null mice. The early onset of AST and ALT rise was not different but later increases were significantly less in null mice coupled with improved survival. Histological evaluation revealed less severe necrosis (limited to a few layers of pericentral cells) in null mice and less DNA degradation. APAP metabolism, as reflected in the kinetics of glutathione depletion, was not affected by the absence of Dnasel. DNA damage activates PARP which consumes NAD+ and NADH (ADP ribosylation) and ATP for DNA repair; the expected loss of NADH and ATP in wild type was not seen in null mice (see Fig.) These provocative findings indicate that in an attempt to repair DNA damage. ATP is severely depleted. A critical event in this process is the release of Dnasel, possibly from damaged ER. The progression of APAP-induced ATP depletion and necrosis from O2 deprived pericentral hepatocytes toward the O2 rich portal tracks depended on the presence of Dnase1, creating a vicious cycle of more necrosis and DNA damage. (See HEPATOLOGY 2006;43:297–305.)

Illustration 2.

Probing the Mind of Mice

Fatigue, lethargy and malaise are debilitating, CNS symptoms of hepatobiliary diseases. Aside from probable neurohumoral communications between the liver and the brain, Kerfoot et al. explored a blood brain barrier/monocyte brain recruitment hypothesis using the bile duct ligated (BDL) mouse model. BDL lead to increased cerebral microvascular VCAM-1 reflecting activation. Circulating monocytes were also activated and exhibited increased TNF production. Intravital microscopy demonstrated increased P-selection dependent rolling of leukocytes in cerebral microvessels (see Fig.). In addition, brain infiltration of activated monocytes but not T cells was increased by BDL. Brain infiltration was blocked by a combination of anti-P-selectin and anti-α4integrin. Irrespective of whether the blood brain barrier is activated by a humoral (cytokines) or circulating cellular mechanism, the consequent infiltration with activated monocytes represents a potentially important new finding. The consequences and relative role of this infiltration of activated monocytes vis-a-vis neurohumoral mechanisms and resident microglial activation remains to be elucidated but the implications extend far beyond nonspecific symptoms. If only we could ask BDL mice if they feel better after inhibition of monocyte infiltration. (See HEPATOLOGY 2006;43:154–162.)

Illustration 3.

If You're Fat, Don't Do That

Obesity is a risk factor for alcoholic liver disease. To assess the possible pathogenetic interactions of ethanol and obesity, Robin et al. examined the effect of 4 consecutive moderate daily doses of 2.5g/kg ethanol on the liver of leptin-deficient ob/ob versus lean mice. Comparable blood ethanol levels were achieved in both. Obese mice developed marked increased hepatocyte apoptosis. Pentoxifylline (PTX) pretreatment decreased elevated ALT levels in ob/ob mice and decreased liver triglyceride and plasma TNF in ob/ob mice given ethanol. Ethanol treatment of ob/ob activated caspase 3 which was blocked by PTX. Ethanol treatment of obese mice was accompanied by decreased basal nuclear p65 NF-kB which was reversed by PTX. This was accompanied by small changes in NF-kB gene expression. These findings suggest that in ob/ob mice ethanol blocks the expected NF-kB response to increased TNF. The difference in response to ethanol in lean versus obese mice could not be explained by parameters of pro or antioxidants or oxidative stress. Overall, the findings are intriguing and generate a number of questions e.g. could obese mice be more susceptible to gut permeabilization and LPS exposure or could their Kupffer cells be primed to respond to LPS (increased TNF). At the same time decreased NF-kB activation (possibly due to increased Hsp70 expression) may sensitize hepatocytes to TNF. It would be important to examine the effect of ethanol on the cytoplasmic assembly of the TNF-receptor complex, explain how caspase 3 is activated in the apparent absence of cytochrome c release, and provide alternative evidence regarding the specificity of the effect of PTX on TNF (e.g. TNF-R1 null ob/ob mice treated with ethanol). Finally, since fatty liver is sensitive to ischemic injury, could the effect of ethanol be due to relative hypoxia? (See HEPATOLOGY 2005;42:1280–1290).

Acceptable Casualties in the War Against Fibrosis

Increasing evidence supports the induction of stellate cell apoptosis as a anti-fibrosis strategy. Activation of NF-kB accompanies stellate cell activation and promotes survival. Therefore, Anan et al. studied the effect of inhibition of NF-kB by blocking proteasomal degradation of IkB. In both a human stellate cell line and primary rat stellate cells, proteasome inhibitors induced apoptosis. Constitutive activation of NF-kB (nuclear p65) was blocked by the proteasome inhibitor, bortezomib (Fig.). A specific NF-kB inhibitor also induced apoptosis. Among the pro-survival Bcl2 family members, A1 expression was decreased by proteasomal inhibition and siRNA silencing of A1 induced apoptosis in these cells. Bortezomib increased expression of Bim, TRAIL-R2, and TRAIL but blocking these did not prevent apoptosis. Treatment of 7-day BDL mice with bortezomib for 4 days (after stellate cell activation) resulted in a marked decrease in activated stellate cells (α-SMA staining), collagen 1α mRNA and increased stellate cell apoptosis (co-localization of TUNEL and α-SMA staining). This study is an elegant demonstration of the potential of inducing stellate cell apoptosis by interfering with the NF-kB pathway as an anti-fibrosis strategy. Among the questions that remain to be answered are: what is the pro-apoptotic stimulus in activated stellate cells, how safe is this strategy (perhaps a more selective targeting of proteasomes or NF-kB in stellate cells will be required) and does this strategy actually decrease collagen deposition or affect bile duct proliferation in the in vivo BDL model? (See HEPATOLOGY 2006;43:335–344.)

Illustration 4.

Construction of Obstruction Without Production of Destruction

Like most biological processes, apoptosis is the net result of promotion and antagonism. Bile duct ligation (BDL) leads to a state of resistance to apoptosis. Osawa et al. studied the mechanisms of anti-apoptosis in BDL mice. BDL mice did not develop massive apoptosis or hemorrhage in the liver after TNF/galactosamine whereas Jo2 (agonistic anti-Fas) still induced moderate hepatocyte apoptosis but no hemorrhage. Using partial BDL, the non-obstructed lobe was not protected demonstrating that the protection was due to local and not systemic factors. TNF did not activate nuclear NF-kB in the obstructed lobe. IkB super-repressor expression sensitized to TNF injury in the non-obstructed lobe but not in the obstructed lobe. BDL lead to activation of AKT which is known to occur in response to bile acid retention and is a protective kinase. Dominant negative AKT overcame the resistance to TNF plus galactosamine or IkB super-repressor. Another finding was selective induction of sphingosine kinase in stellate cells from the obstructed lobe. Sphingosine kinase KO mice were no longer resistant to apoptosis and hemorrhage from TNF and galactosamine or hemorrhage from Jo2. MMP-2 and MMP-9 were activated in both lobes after all pro-apoptotic treatments but TIMP-1 (MMP inhibitor) expression was increased in the obstructed lobe; this was suppressed in KO mice. The key finding is that blocking NF-kB activation did not sensitize hepatocytes to TNF-induced apoptosis unless AKT was inhibited. Why was NF-kB not activated by TNF in obstructed hepatocytes and what are the critical targets of AKT in this context? Anti-hemorrhagic effect of BDL was blocked in sphingosine kinase KO mice independent of AKT. The results suggest that hemorrhage and aggravated apoptosis is due to the action of MMPs. Sphingosine kinase dependent production of TIMP-1 by stellate cells in BDL inhibits the contribution of MMP to hemorrhage and apoptosis. The implication is that these anti-apoptotic effects prevent acute liver failure and allow chronic liver injury and fibrosis to progress. (See HEPATOLOGY 2005;42:1320–1328.)