Glucocorticoids Have Opposing Effects on Liver Fibrosis in Hepatic Stellate and Immune Cells

Abstract

Liver fibrosis is a reversible wound-healing process that is protective in the short term, but prolonged fibrotic responses lead to excessive accumulation of extracellular matrix components that suppresses hepatocyte regeneration, resulting in permanent liver damage. Upon liver damage, nonparenchymal cells including immune cells and hepatic stellate cells (HSCs) have crucial roles in the progression and regression of liver fibrosis. Here, we report differential roles of the glucocorticoid receptor (GR), acting in immune cells and HSCs, in liver fibrosis. In the carbon tetrachloride hepatotoxin-induced fibrosis model, both steroidal and nonsteroidal GR ligands suppressed expression of fibrotic genes and decreased extracellular matrix deposition but also inhibited immune cell infiltration and exacerbated liver injury. These counteracting effects of GR ligands were dissociated in mice with conditional GR knockout in immune cells (GR(LysM)) or HSC (GR(hGFAP)): the impacts of dexamethasone on immune cell infiltration and liver injury were totally blunted in GR(LysM) mice, whereas the suppression of fibrotic gene expression was diminished in GR(hGFAP) mice. The effect of GR activation in HSC was further confirmed in the LX-2 HSC cell line, in which antifibrotic effects were mediated by GR ligand inhibition of Sma and mad-related protein 3 (SMAD3) expression. We conclude that GR has differential roles in immune cells and HSCs to modulate liver injury and liver fibrosis. Specific activation of HSC-GR without alteration of GR activity in immune cells provides a potential therapeutic approach to treatment of hepatic fibrosis.

Figure 1.. GR is highly expressed in KCs and HSC.

A, C57BL6/J wild-type livers were stained with anti-GR antibody (×200). Arrowheads indicate GR-positive signals in nonparenchymal cells. B, The human liver sections stained with 2 kinds of GR antibody (CAB010435 and HPA004248) provided by The Human Protein Atlas (available at http://www.proteinatlas.org). Arrowhead indicate GR-positive signals in nonparenchymal cells. C, Comparison of GR expression level in isolated HCs, KCs and HSCs. Purity of each cell fraction was confirmed by the expression levels of albumin (Alb), macrophage antigen (CD68), and Col1α1. Asterisks denote statistically significant differences (one-way ANOVA; *, P < .05; ***, P < .001). Normal livers were double-stained with (D) anti-GR antibody (red) and F4/80 antibody (green) or (E) anti-GR antibody (red) and anti-GFAP antibody (green). Nuclei were visualized by DAPI staining (blue). Arrowheads indicate colocalization of GR and cell-specific markers. CV, central vein.

Figure 2.. DEX attenuates CCl4-induced hepatic fibrosis but exacerbates liver injury.

A, Wild-type mice were treated with CCl₄ plus vehicle (CCl₄+PBS, n = 7) or CCl₄ plus water-soluble form of DEX (DEX-NaPO₄; CCl₄+DEX, n = 7). Liver sections were stained with hematoxylin/eosin (H&E) (×100, upper) and Sirius red (×40, lower). Arrowheads indicate areas of infiltrating mononuclear cells. B, The number of infiltrated cells were counted. C, The size of fibrotic area was quantified. D, Serum ALT level of DEX-treated mice. E, Expressions of fibrotic genes in the presence of CCl₄ and/or DEX. Asterisks denote statistically significant differences (Student's t test; *, P < .05; **, P < .01; ***, P < .005; †, Cohen's d > 0.8).

Figure 3.. GR modulator, CpdA, has comparable effects with DEX.

A, Wild-type mice were treated with CCl₄ plus vehicle (CCl₄+PBS, n = 5) or CCl₄ plus CpdA (CCl₄+CpdA, n = 5). Liver sections were stained with hematoxylin/eosin (H&E) (×100, upper) and Sirius red (×40, lower). Arrowheads indicate areas of infiltrating mononuclear cells. B, The number of infiltrated cells were counted. C, The percentage of fibrotic area was calculated. D, Serum ALT level of CpdA-treated mice. E, Fibrotic gene expression in CpdA-treated animals. Asterisks denote statistically significant differences (Student's t test; *, P < .05; **, P < .01; ***, P < .005; †, Cohen's d > 0.8).

Figure 4.. Immune cell-specific GR deletion reversed DEX effect on immune cell infiltration and liver injury.

A, CCl₄-tretaed immune cell-specific GR knockout mice (GR^LysM) and control littermates (GR^F/F) were administered with vehicle (PBS, n = 3 per each genotype) or DEX-NaPO₄ (DEX, n = 3 per each genotype). Representative liver sections stained with hematoxylin/eosin (H&E) (×100, upper) and Sirius red (×40, lower) are shown. Arrowhead indicates areas of infiltrating mononuclear cells. The number of infiltrated mononuclear cells (B), size of fibrotic area (C), and serum ALT level (D) were analyzed. E, Fibrotic gene expressions were quantified by Q-PCR analysis. Asterisks denote statistically significant differences (Student's t test; *, P < .05; **, P < .01; ***, P < .005; †, Cohen's d > 0.8).

Figure 5.. HSC-specific GR deletion abolishes DEX effect on fibrotic gene expression.

A, HSC-specific GR knockout mice (GR^hGFAP) and control littermates (GR^F/F) were injected with CCl₄ in combination with vehicle (PBS, n = 3 per each genotype) or DEX-NaPO₄ (DEX, n = 3 per each genotype). Representative liver sections stained with hematoxylin/eosin (H&E) (×100, upper) and Sirius red (×40, lower) are shown. Arrowhead areas of infiltrating mononuclear cells. B, Infiltrated mononuclear cells were counted. C, The size of fibrotic area were quantified. D, Serum ALT level was analyzed. E, The changes of fibrotic gene expression were determined by Q-PCR analysis. Asterisks denote statistically significant differences (Student's t test; *, P < .05; **, P < .01; †, Cohen's d > 0.8).

Figure 6.. GR ligands suppress fibrotic gene expression in HSC.

LX-2 cells were pretreated with (A) Steroidal GR agonists (DEX; water-soluble DEX, DEX-NaPO₄; prednisolone; budesonide; 2μM) and (B) GR modulator (CpdA; 2μM or 4μM) were pretreated. They were then treated with recombinant human TGFβ1 (5 ng/mL) for additional 24 hours, and Q-PCR analysis of fibrotic genes was carried out. Asterisks denote statistically significant differences compared with vehicle-treated sample (Student's t test; *, P < .05; **, P < .01; ***, P < .001). C, LX-2 cells were transfected with NT siRNA and siGR. After 12 hours, DEX (2μM) and TGFβ1 (5 ng/mL) were treated for 24 hours, and the fibrotic gene expression was analyzed by Q-PCR. In the presence of TGFβ1, the relative changes of fibrotic gene expression by DEX were denoted (%). Asterisks denote statistically significant differences compared with vehicle-treated sample (Student's t test; *, P < .05; **, P < .01; ***, P < .001).

Figure 7.. Inhibition of SMAD3 expression mediates antifibrotic effect of GR.

A, LX-2 cells were pretreated with various GR ligands (2μM) for 2 hours then recombinant human (rh) TGFβ1 (5 ng/mL) was added. After 24 hours, expressions of COL1α1 pre-mRNA level were analyzed by Q-PCR. B, SMAD2 and SMAD3 expressions were determined by Q-PCR. C, LX2 cells were transfected with NT siRNA or siGR. After 12 hours, DEX and TGFβ1 were treated for additional 24 hours, and the SMAD2/3 expression was analyzed by Q-PCR. The relative changes by DEX treatment were denoted (%). Asterisks denote statistically significant differences compared with vehicle-treated sample (Student's t test; *, P < .05; **, P < .01; ***, P < .001). D, Summary of differential roles of GR in immune cell and HSC on liver fibrosis.