Insights into molecular mechanisms contributing to individual susceptibility to steatohepatitis: Lessons learned from mouse models


  • Potential conflict of interest: Nothing to report.

  • Supported by Austrian Wirtschaftsservice (IMGuS) and the Austrian Genome Programme GEN-AU.

Snider NT, Weerasinghe SV, Singla A, Leonard JM, Hanada S, Andrews PC, et al. Energy determinants GAPDH and NDPK act as genetic modifiers for hepatocyte inclusion formation. J Cell Biol 2011;195: 217-229. (Reprinted with permission.)


Genetic factors impact liver injury susceptibility and disease progression. Prominent histological features of some chronic human liver diseases are hepatocyte ballooning and Mallory-Denk bodies. In mice, these features are induced by 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) in a strain-dependent manner, with the C57BL and C3H strains showing high and low susceptibility, respectively. To identify modifiers of DDC-induced liver injury, we compared C57BL and C3H mice using proteomic, biochemical, and cell biological tools. DDC elevated reactive oxygen species (ROS) and oxidative stress enzymes preferentially in C57BL livers and isolated hepatocytes. C57BL livers and hepatocytes also manifested significant down-regulation, aggregation, and nuclear translocation of glyceraldehyde 3-phosphate dehydrogenase (GAPDH). GAPDH knockdown depleted bioenergetic and antioxidant enzymes and elevated hepatocyte ROS, whereas GAPDH overexpression decreased hepatocyte ROS. On the other hand, C3H livers had higher expression and activity of the energy-generating nucleoside-diphosphate kinase (NDPK), and knockdown of hepatocyte NDPK augmented DDC-induced ROS formation. Consistent with these findings, cirrhotic, but not normal, human livers contained GAPDH aggregates and NDPK complexes. We propose that GAPDH and NDPK are genetic modifiers of murine DDC-induced liver injury and potentially human liver disease.


A variety of diseases lead to formation of cytoplasmic inclusions in hepatocytes such as Mallory-Denk bodies (MDBs), intracytoplasmic hyaline bodies, pale bodies, ground-glass inclusions, glycogen bodies, and alpha-1-antitrypsin inclusions. Among these, MDBs are the most frequently observed inclusions and are characteristic morphological features of alcoholic (ASH) and nonalcoholic steatohepatitis (NASH).1 The chemical composition of MDBs was characterized based on several studies, and keratins, sequestosome 1/p62, ubiquitin, and heat shock proteins 70 and 25 were found to be the major proteins. Other than the composition and molecular structure of MDBs, the biological and clinical significance of MDBs is still an open issue. Although a variety of cellular mechanisms were found altered in association with MDB-formation (e.g., increased oxidative stress, misfolding and aggregation of proteins, protein cross-linking, deficiencies of the protein degradation machinery), a causal relationship is still not unequivocally established.1 Furthermore, an important open question is why only a subpopulation of patients with similar risk factors develops steatohepatitis, and why the formation of MDBs is a frequent but not obligatory feature of hepatocyte damage in steatohepatitis. For instance, only 20% of heavy drinkers develop ASH, whereas 20% of type II diabetic and 50% of obese type II diabetic patients develop NASH. These variations in disease manifestation might, at least partly, be attributed to genetic risk factors, because there is concordance in monozygotic twins, dependence on ethnic origin (Hispanics are more susceptible than Caucasians and African-Americans), and impact of sex.2-4

Snider et al.5 now provide important new insight into the impact of genetic background on the pathophysiology of hepatocyte injury in association with MDB formation. They used two different inbred mouse strains that revealed different susceptibility to MDB formation upon long-term intoxication with 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) as a model system. In a previous study the same group had already demonstrated that the prevalence of key phenotypes of hepatocyte injury in steatohepatitis (i.e., hepatocyte ballooning, MDB formation, steatosis, and apoptosis) markedly vary in C57BL/6, FVB/N, Balb/cAnN, C3H/He, and 129X1/Sv mouse strains.6 For instance, ballooning was most prominent in C57BL/6, whereas C3H/He had the lowest ballooning scores. Using these two strains with the most striking differences in their response to DDC-treatment, Snider et al. first performed a screening approach with 2D differential in-gel electrophoresis to obtain an overview of differences in protein expression. Differentially expressed proteins were identified by mass spectrometry analysis, then validated, and could be classified into three different groups (i.e., protein processing, energy metabolism, oxidative stress groups). Some of these proteins had more than 10-fold different expression levels in these two strains even without toxic challenge. For instance, nucleoside diphosphate kinase A/B (NDPK) had a 12-fold higher expression in C3H/He mice and the difference was still approximately 5-fold after DDC feeding. Enzymes related to the oxidative stress response, such as carbonyl reductase 3 (CBR3), or glutathione-S-transferase were more highly expressed in C3H/He mice on normal diet but were markedly increased in DDC-fed C57BL/6 mice. These data indicated profound alterations in basic biological processes in these mouse strains under basal conditions that could lay the foundation for different responses under disease conditions. Interestingly, some of these alterations (e.g., NDPK) were not seen at the messenger RNA (mRNA) expression level but only became evident after using a proteomics approach. Furthermore, in addition to the striking alteration of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression, aggregation and nuclear translocation of this enzyme were observed. NDPH and GAPDH are of particular interest in the context of MDB pathogenesis. NDPK catalyzes the transfer of the terminal phosphate of nucleoside triphosphates to nucleoside diphosphates, thereby being important for the synthesis of GTP, CTP, and UTP. NDPK (also called Nm23) has been identified as a metastasis suppressor whose function is regulated by oxidation and reduction at its Cys109.7 GAPDH is best known as a glycolytic “housekeeping” cytoplasmic enzyme. However, it appears to be involved in several additional processes such as DNA repair, transfer RNA (tRNA) export, membrane fusion, cytoskeletal dynamics, and cell death.8, 9 These different functions are, at least in part, regulated by oxidative stress, resulting in posttranslational modifications, oligomerization, aggregation, and nuclear translocation of GAPDH.

To further understand the biological relevance of the strain-specific alterations of NDPK and GAPDH, a series of in vitro and ex vivo experiments were performed by Snider et al. Using the reactive oxygen species (ROS)-sensitive fluorescent probe CM-H2DCFDA they demonstrated that DDC resulted in higher ROS levels in cultured C57BL/6 than in C3H/He hepatocytes. This increase in ROS was paralleled by an increase in CBR3 immunoreactivity, resembling the in vivo situation observed in livers of mice. Furthermore, DDC treatment of hepatocytes isolated from C57BL/6 mice resulted in a dose-dependent reduction of cytoplasmic GAPDH and nuclear translocation which could be inhibited by pioglitazone. Using small interfering RNA (siRNA)-mediated knockdown, GAPDH was identified as an upstream regulator of NDPK and other enzymes involved in antioxidant responses. Consequently, knockdown of GAPDH or NDPK resulted in increased ROS formation in hepatocytes. Interestingly, knockdown of NDPK resulted also in a decrease of GAPDH, suggesting a coregulation of these two enzymes. The effects of GAPDH knockdown were not restricted to antioxidant enzymes but also affected metabolic functions, e.g., by down-regulation of fumarylacetoacetate hydrolase.

The work of Snider et al. has clearly shown the impact of genetic background on the hepatotoxicity of DDC. Although no genetic modifiers as such (e.g., allelic heterogeneity of the differently expressed genes) have been identified in the two mouse strains investigated, the affected cellular mechanisms shed some light on genes that might also influence an individual's susceptibility to develop steatohepatitis. The presented data highlight the role of oxidative stress in DDC toxicity, which could be the common pathophysiological denominator between DDC-induced liver injury and ASH.10 In particular, Snider et al. demonstrate that the increased antioxidant response in C57BL/6 hepatocytes is a consequence of increased oxidative stress and not a sign of increased stress resistance. Furthermore, they established a regulatory antioxidant network involving GAPDH and NDPK as key elements. Moreover, major changes were observed in energy metabolism, which underlines the hypothesis that energy-dependent protein-folding and -degradation are critical factors for MDB formation. Another interesting finding was that the DDC-induced alterations of GAPDH could be prevented by pioglitazone (Fig. 1). This could be a novel mechanistic explanation for the improved liver histology in NASH patients treated with pioglitazone.11

Figure 1.

Diagram of pathophysiological alterations that are associated with susceptibility to steatohepatitis in C57BL/6 mice. ROS leads to down-regulation, aggregation, and nuclear translocation of GAPDH. This results in profound alterations in energy metabolism and redox reactions (redox crisis), which impairs protein refolding and degradation machinery, thereby favoring the formation of MDBs. ROS, reactive oxygen species; GAPDH, glyceraldehydes 3-phosphate dehydrogenase, NDPK, nucleoside-diphosphate kinase; FAH, fumarylacetoacetate hydrolase; CA3, carbonic anhydrase 3; PRDX6, peroxiredoxin 6; GST, glutathione S transferase; CBR3, carbonyl reductase 3; MDBs, Mallory-Denk bodies. Adapted from Snider et al., 2011.

Although the above-described mechanisms were found to be differently affected in mouse strains either susceptible or resistant to develop features of steatohepatitis after DDC intoxication, a direct causal relationship of the findings to steatohepatitis and MDB formation remains to be shown. On the one hand, the observation that nuclear translocation and aggregation of GAPDH and NDPK was also found in explants of human livers with endstage alcoholic liver disease is further evidence for the general validity and the close relationship of these alterations to the pathophysiology of the disease. On the other hand, cirrhosis typically dominates steatohepatitis in explanted livers as the leading phenotype; hence, a direct casual relationship to MDB formation remains to be investigated in human livers as well.

Very important general take-home messages can be drawn in addition to the relevance of the study to better understand the pathophysiology of steatohepatitis. Above all, the impact of the genetic background on mouse models is often underestimated in its capacity to explain discrepancies in the findings obtained by different research laboratories.12 Unfortunately, the genetic background of mice used in such studies is often insufficiently described in publications.

Second, this study provides a compelling example that important molecular alterations might not become evident by gene-expression profiling but might rather require detailed analysis of proteins including their intracellular localization and interaction partners. This could well explain why the important role of GAPDH in steatohepatitis-associated liver injury has so far been overlooked.