Development of nonalcoholic steatohepatitis in insulin-resistant liver-specific S503A carcinoembryonic antigen-related cell adhesion molecule 1 mutant mice

Abstract

Background & aims: Liver-specific inactivation of carcinoembryonic antigen-related cell adhesion molecule 1 causes hyperinsulinemia and insulin resistance, which result from impaired insulin clearance, in liver-specific S503A carcinoembryonic antigen-related cell adhesion molecule 1 mutant mice (L-SACC1). These mice also develop steatosis. Because hepatic fat accumulation precedes hepatitis, lipid peroxidation, and apoptosis in the pathogenesis of nonalcoholic steatohepatitis (NASH), we investigated whether a high-fat diet, by causing inflammation, is sufficient to induce hepatitis and other features of NASH in L-SACC1 mice.

Methods: L-SACC1 and wild-type mice were placed on a high-fat diet for 3 months, then several biochemical and histologic analyses were performed to investigate the NASH phenotype.

Results: A high-fat diet caused hepatic macrosteatosis and hepatitis, characterized by increased hepatic tumor necrosis factor alpha levels and activation of the NF-kappaB pathway in L-SACC1 but not in wild-type mice. The high-fat diet also induced necrosis and apoptosis in the livers of the L-SACC1 mice. Insulin resistance in L-SACC1 fed a high-fat diet increased the hepatic procollagen protein level, suggesting a role in the development of fibrosis.

Conclusions: A high-fat diet induces key features of human NASH in insulin-resistant L-SACC1 mice, validating this model as a tool to study the molecular mechanisms of NASH.

Figure 1

Northern analysis of hepatic enzymes. Liver was extracted from fed WT and L-SACC1 mice, which had been RD- or HF-fed for 3 months, and mRNA was purified and analyzed sequentially for (A) Cpt-1, Pdk4, and Gapdh levels. (B) Another aliquot was analyzed for G-6-Pase and Gapdh mRNA levels. C) and D) Other aliquots were analyzed for Srebp1c and Pepck, respectively, before reprobing with Gapdh mRNA levels. Bands were quantitated by densitometry and presented in the graph as arbitrary units. n>5 from each feeding group of mice. Values are mean±S.E. *P<0.05 HF versus RD.

Figure 2

Hepatic lipid metabolism. Liver lysates (A) from >3 mice from each of the four feeding groups (RD- and HF-WT, and RD-and HF-L-SACC1) were analyzed by immunoblotting (IB) with α-FAS, α-PPARγ and α-NPC-1 antibodies followed by reprobing (reIB) with α-actin or α-GAPDH to normalize for the amount of protein loaded. (B) 10μg serum proteins from >5 mice per feeding group were analyzed by Western blotting with α-ApoB antibody to detect ApoB100 (~500kDa) and ApoB48 (~200kDa), which were quantitated by densitometry and presented in the bar graph as arbitrary units. Values are mean±S.E. *P<0.05 HF versus RD, ^†P< 0.05 versus WT-RD. (C) Liver sections from >5 mice of each of WT-RD (block 1), WT-HF (block 2), L-SACC1-RD (block 3) and L-SACC1-HF (block 4) were analyzed by H&E staining for lipid accumulation.

Figure 3

Characterization of hepatic inflammation. (A) Liver sections from >5 mice per feeding group, were analyzed by H&E staining for inflammatory cell infiltration. (B) Hepatic Tnfα mRNA levels from >3 mice per feeding group were determined by Northern analysis, as in Figure 1. Values are mean±S.E. *P<0.05 HF versus RD, ^†P< 0.05 versus WT-RD. (C) Liver lysates from >3 mice per feeding group, were analyzed by Western blotting (IB) with α-phospho-NF-κB (upper gel) and reprobing (reIB) with α-NF-κB (lower gel). Activation of PKCζ was similarly analyzed for the amount of phosphorylated (upper gel) relative to protein level of PKCζ (lower gel). Bands were quantitated by densitometry and presented in the graph as arbitrary units.

Figure 4

Immunohistochemical analysis of inflammation. Liver sections were analyzed with (A) FITC-conjugated anti-F4/80 antibody and (B) FITC-conjugated anti-CD3 and anti-CD4 antibodies. Experiments were repeated on 3 sections per mouse. Counts are mean positively stained cells per visible field±S.E. *P<0.05 HF versus RD.

Figure 5

Nitroso-redox evaluation. (A) Liver lysates from >3 mice per feeding group were analyzed by immunoblotting (IB) sequentially with α-PPARα, α-UCP2 and α-PGC1α antibodies prior to reprobing (reIB) with α-actin (Upper group of gels). Another aliquot was similarly analyzed with α-CYP2E1 antibody and α-actin (lower group of gels). Bands were quantitated by densitometry and presented in the bar graph as arbitrary units. (B) Liver homogenates from >5 mice per feeding group were assayed for TBARS/MDA and GSH, and serum was assayed for nitrite levels. Values are mean±S.E. *P<0.05 HF versus RD, ^†P<0.05 versus WT-RD.

Figure 6

Analysis of cell injury. (A) Liver lysates from >5 mice were analyzed by immunoblotting (IB) with α-PARP prior to reprobing with α-Tubulin. (B) Liver sections from all four feeding groups were analyzed by TUNEL (block 1–4) and H&E staining (block 5). Injured cells are emphasized by arrowheads (blocks 4 &5).

Figure 7

Analysis of fibrosis. (A) Liver lysates from >3 mice were analyzed by immunoblotting (IB) with α-procollagen antibody prior to reprobing (reIB) with α-tubulin. (B) Liver sections from >5 mice of each feeding group were analyzed by Sirius red staining.