SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

Although nonalcoholic steatohepatitis (NASH) is associated with hypercholesterolemia, the underlying mechanisms of this association have not been clarified. We aimed to elucidate the precise role of cholesterol in the pathophysiology of NASH. C57BL/6 mice were fed a control, high-cholesterol (HC), methionine-choline-deficient (MCD), or MCD+HC diet for 12 weeks or a control, HC, high-fat (HF), or HF+HC diet for 24 weeks. Increased cholesterol intake accelerated liver fibrosis in both the mouse models without affecting the degree of hepatocellular injury or Kupffer cell activation. The major causes of the accelerated liver fibrosis involved free cholesterol (FC) accumulation in hepatic stellate cells (HSCs), which increased Toll-like receptor 4 protein (TLR4) levels through suppression of the endosomal-lysosomal degradation pathway of TLR4, and thereby sensitized the cells to transforming growth factor (TGF)β-induced activation by down-regulating the expression of bone morphogenetic protein and activin membrane-bound inhibitor. Mammalian-cell cholesterol levels are regulated by way of a feedback mechanism mediated by sterol regulatory element-binding protein 2 (SREBP2), maintaining cellular cholesterol homeostasis. Nevertheless, HSCs were sensitive to FC accumulation because the high intracellular expression ratio of SREBP cleavage-activating protein (Scap) to insulin-induced gene (Insig) disrupted the SREBP2-mediated feedback regulation of cholesterol homeostasis in these cells. HSC activation subsequently enhanced the disruption of the feedback system by Insig-1 down-regulation. In addition, the suppression of peroxisome proliferator-activated receptor γ signaling accompanying HSC activation enhanced both SREBP2 and microRNA-33a signaling. Consequently, FC accumulation in HSCs increased and further sensitized these cells to TGFβ-induced activation in a vicious cycle, leading to exaggerated liver fibrosis in NASH. Conclusion: These characteristic mechanisms of FC accumulation in HSCs are potential targets to treat liver fibrosis in liver diseases including NASH. (Hepatology 2014;58:154–169)

Abbreviations
ABCA1

adenosine triphosphate-binding cassette A1

ALT

alanine aminotransferase

Bambi

bone morphogenetic protein and activin membrane-bound inhibitor

CCl4

carbon tetrachloride

CE

cholesterol ester

COPII

coat protein complex II

ER

endoplasmic reticulum

FBS

fetal bovine serum

FC

free cholesterol

HC

high cholesterol

HF

high fat

HMGCR

3-hydroxy-3-methyl-glutaryl-CoA reductase

HSC

hepatic stellate cell

ICAM-1

intercellular adhesion molecule-1

Insig

insulin-induced gene

LDLR

low-density lipoprotein receptor

LPS

lipopolysaccharide

Mβ CD

methyl-β-cyclodextrin

MCD

methionine-choline deficient

mRNA

messenger RNA

NASH

nonalcoholic steatohepatitis

NPC1

Niemann-Pick C1

PCR

polymerase chain reaction

PPAR

peroxisome proliferator-activated receptor

Scap

SREBP cleavage-activating protein

siRNA

small interfering RNA

SMA

smooth muscle actin

SREBP

sterol regulatory element-binding protein

TGF

transforming growth factor

TLR4

Toll-like receptor 4

TNF

tumor necrosis factor

TUNEL

terminal deoxynucleotidyl transferase-mediated deoxyuridine nick-end labeling

VCAM-1

vascular cell adhesion molecule-1.

Nonalcoholic steatohepatitis (NASH) is a progressive disease that can cause cirrhosis or liver-related complications.[1] It very often accompanies lifestyle diseases including hypercholesterolemia. Several studies have shown that statins and ezetimibe (cholesterol-lowering agents) improve liver fibrosis in patients with NASH.[2] Furthermore, we have recently reported that free cholesterol (FC) accumulation in hepatic stellate cells (HSCs) plays an important role in the pathogenesis of liver fibrosis.[3] These results drew our attention to the role of cholesterol in the pathogenesis of liver fibrosis in NASH.

Cholesterol homeostasis is tightly regulated by way of a feedback system mediated by sterol regulatory element-binding protein (SREBP)2.[4, 5] The low-density lipoprotein receptor (LDLR) and 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMGCR), which play important roles in maintaining cholesterol uptake and synthesis, respectively, are predominantly regulated by SREBP2.[6] Nascent SREBP2 localizes to the endoplasmic reticulum (ER) membrane and forms tight complexes with SREBP cleavage-activating protein (Scap), a membrane-embedded escort protein.[7] When membrane cholesterol levels are low, the SREBP2-Scap complex is incorporated into the coat protein complex II (COPII)-coated vesicles.[6, 8] Consequently, SREBP2 translocates to the nucleus and activates transcription of several target genes involved in the biosynthesis and uptake of cholesterol.[6] When excess cholesterol accumulates in the ER membranes, it changes Scap to an alternate conformation, allowing it to bind to resident ER proteins, insulin-induced gene (Insig)-1, and Insig-2.[9] This binding precludes the binding of COPII. Consequently, the SREBP2-Scap complex remains in the ER, transcription of the target genes declines, and cholesterol synthesis and uptake fall.[4, 6]

Furthermore, recent studies have shown that the primary transcript of SREBP2 also encodes miR-33a, a microRNA that regulates cholesterol metabolism by way of factors such as adenosine triphosphate-binding cassette A1 (ABCA1) and Niemann-Pick C1 (NPC1), suggesting transcriptional regulation by SREBF2 modulates the cellular capacity for producing not only an active transcription factor but also the expression of miR-33a.[10]

By studying two mouse models of NASH, we attempted to clarify the precise role of cholesterol in the pathophysiology of NASH. As we found that the major causes of the exacerbation of liver fibrosis in NASH involved FC accumulation in HSCs, we investigated the underlying mechanisms of FC accumulation in HSCs and its role in the pathogenesis of NASH.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

Please refer to the Supporting Materials and Methods for more detailed descriptions.

Reagents

Reagents were obtained as follows: low density lipoprotein (LDL), methyl-β-cyclodextrin (MβCD)/cholesterol complex, lipopolysaccharide (LPS), chloroquine, and MG-132 were from Sigma (St. Louis, MO). 25-HC was from Wako Pure Chemical Industries (Osaka, Japan). Transforming growth factor beta (TGFβ) was from R&D Systems (Minneapolis, MN). Peroxisome proliferator-activated receptor gamma (PPARγ)-small interfering RNA (siRNA), SREBP2-siRNA, LDLR-siRNA, Scap-siRNA, Insig-1-siRNA, bone morphogenetic protein and activin membrane-bound inhibitor (Bambi)-siRNA, and control-siRNA were from Invitrogen (Carlsbad, CA). Anti-miR33a, pre-miR33a, and control-miR33a were from Ambion (Austin, TX).

Animal Studies

Nine-week-old male C57BL/6J mice (CLEA Japan, Tokyo, Japan) were fed a CE-2 (control; CLEA Japan), CE-2 with 1% cholesterol (HC), methionine-choline-deficient (MCD; Cat. No. 960439; ICN, Aurora, OH), or MCD with 1% cholesterol (MCD+HC) diet for 12 weeks. As another animal model of NASH, 9-week-old male C57BL/6J mice were also fed a CE-2, HC, high-fat (HF; prepared by CLEA Japan according to the #101447 composition of Dyets, Bethlehem, PA), or HF with 1% cholesterol (HF+HC) diet for 24 weeks. In the same way, 7-8-week-old C57BL/6 Toll-like receptor (TLR)4-deficient mice (Oriental BioService, Kyoto, Japan) were fed the control, HC, MCD, or MCD+HC diets for 8 weeks or the control, HC, HF, or HF+HC diets for 20 weeks. All animals received humane care in compliance with the criteria outlined in the “Guide for the Care and Use of Laboratory Animals,” prepared by the US National Academy of Sciences and published by the US National Institutes of Health.

HSC Isolation and Cell Culture

Wild-type or TLR4-deficient HSCs were isolated from the livers of mice as described.[3] We cultured HSCs on uncoated 6-well plastic tissue culture dishes in serum-depleted Dulbecco's modified Eagle's medium (DMEM), DMEM containing 1% or 10% fetal bovine serum (FBS), and used them as nonpassaged primary cultures or cultures at passage 3-6.

Statistical Analysis

All data are expressed as means (standard error of the mean [SEM]). Statistical analyses were performed using the unpaired Student t test or one-way analysis of variance (ANOVA) (P < 0.05 was considered significant). When the ANOVA analyses were applied, differences in mean values among groups were examined by Fisher's multiple comparison test.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information
Increased Cholesterol Intake Accelerates Liver Fibrosis in NASH Without Affecting the Degree of Hepatocellular Injury or Macrophage Recruitment or Activation

Compared with the livers of the MCD diet-fed mice, the livers of the MCD+HC diet-fed mice showed markedly increased centrizonal fibrosis (Supporting Fig. 1A-C). As observed in the MCD diet-induced NASH model, the extent of fibrosis was significantly enhanced in the livers of the HF+HC diet-fed mice, compared with the HF diet-fed mice (Supporting Fig. 1D-F).

HC diet feeding alone was not sufficient to cause liver fibrosis over 12 and 24 weeks (Supporting Fig. 1). In addition, increased intake of cholesterol did not significantly impact hepatocellular damage in the two mouse models of NASH (Supporting Fig. 2). There was no impact on the hepatic messenger RNA (mRNA) levels of Cyp27a1 or on the hepatic content of mitochondrial FC (Supporting Fig. 3).

Similarly, the increased cholesterol intake did not increase macrophage recruitment or activation in either of the two mouse models of NASH (Supporting Fig. 4). Neither did the increased cholesterol intake induce the formation of hepatic macrophage foam cells or cause liver inflammation in these mouse models (Supporting Figs. 1A,D, 5A). In Kupffer cells, there was also no impact on the mRNA levels of Cyp27a1 or on the cholesterol content of both the mitochondria and late endosomes/lysosomes (Supporting Fig. 5B-D). Furthermore, the increased cholesterol intake significantly exaggerated liver fibrosis in Kupffer cell-depleted mice with NASH (Supporting Fig. 6).

FC Accumulation in HSCs Is Enhanced in NASH and Up-Regulates TLR4 Protein Expression and Down-Regulates Bambi mRNA Expression in HSCs

HC, MCD, and HF diet feeding significantly increased FC levels in HSCs compared with the corresponding control diet feeding (Supporting Fig. 7A,D). Further, FC levels were significantly higher in HSCs from the MCD+HC and HF+HC diet-fed groups than in those from the other corresponding groups (Supporting Fig. 7A,D).

The mRNA expression levels of Bambi, the TGFβ pseudoreceptor, were significantly lower in HSCs from the HC, MCD, and HF diet-fed groups than in those from the corresponding control diet-fed groups and in HSCs from the MCD+HC and HF+HC diet-fed groups than in those from the other corresponding groups (Supporting Fig. 7B,E).

HC, MCD, and HF diet feeding increased the amount of TLR4 protein expressed in HSCs. In addition, HSCs from the MCD+HC and HF+HC diet-fed groups showed higher TLR4 protein expression than those from the other corresponding groups (Supporting Fig. 7C,F). No significant difference was observed in the mRNA expression levels of TLR4 among the corresponding groups (Supporting Fig. 7C,F).

HSC Activation in NASH Down-Regulates PPARγ Expression and Enhances Both SREBP2 and miR-33a Signaling; Increased Cholesterol Intake Intensifies These Effects

As noted in the whole livers, the mRNA expression levels of collagen 1α1, collagen 1α2, and α smooth muscle actin (αSMA) were significantly increased in HSCs from the MCD and HF diet-fed groups compared with the corresponding control diet-fed groups. These increases were significantly enhanced by the increased intake of cholesterol (Fig. 1A,D).

image

Figure 1. Down-regulated PPARγ expression and enhanced SREBP2 and miR-33a signaling after HSC activation in the two mouse models of NASH. C57BL/6 mice (9 weeks old, male; n = 6-9/group) were fed (A-C) the control, HC, MCD, or MCD+HC diet for 12 weeks or (D-F) the control, HC, HF, or HF+HC diet for 24 weeks. (A,D) Quantification of collagen 1α1, collagen 1α2, αSMA, PPARγ1, and SREBP2 mRNA in HSCs isolated from the mice in each group. **P < 0.01 and *P < 0.05, compared with the control diet group. (B,E) Total and nuclear expression of PPARγ and SREBP2 protein in HSCs isolated from the mice in each group. The relative protein levels are indicated below the corresponding bands. (C,F) Quantification of LDLR and HMGCR mRNA, and miR-33a in HSCs isolated from the mice in each group. **P < 0.01 and *P < 0.05, compared with the control diet group. All data are expressed as means (SEM).

Download figure to PowerPoint

The mRNA expression levels of PPARγ1 in HSCs were significantly lower in the MCD and HF diet-fed groups than in the corresponding control diet-fed groups. In addition, these decreases were significantly enhanced by the increased intake of cholesterol (Fig. 1A,D). Contrarily, the mRNA expression levels of SREBP2 were significantly higher in HSCs from the MCD and HF diet-fed groups than in those from the corresponding control diet-fed groups, and these increases were significantly enhanced by the increased intake of cholesterol (Fig. 1A,D).

The total and nuclear protein levels of PPARγ were lower in HSCs from the MCD and HF diet-fed groups than in those from the corresponding control diet-fed groups and these decreases were significantly enhanced by the increased intake of cholesterol (Fig. 1B,E). Meanwhile, the levels of the nuclear form of SREBP2 were significantly higher in HSCs from the MCD and HF diet-fed groups than in those from the corresponding control diet-fed groups. Furthermore, these increases were significantly enhanced by the increased intake of cholesterol (Fig. 1B,E).

Similar to SREBP2 expression, the expression levels of LDLR and miR-33a in HSCs were significantly higher in the MCD and HF diet-fed groups than in the corresponding control diet-fed groups. These increases were significantly enhanced by the increased intake of cholesterol (Fig. 1C,F).

In Vitro HSC Activation Down-Regulates PPARγ Signaling, Which Enhances SREBP2 and miR-33a Signaling

The total and nuclear forms of PPARγ were abundant in day 1 (quiescent) HSCs but declined in day 3 and 5 (activating) and day 7 (activated) HSCs (Fig. 2A). Meanwhile, the nuclear form of SREBP2 was scarce in day 1 HSCs, and its expression increased at days 3 and 5, and day 7 HSCs (Fig. 2A). Correspondingly, the PPARγ1 and SREBP2 mRNA expression levels were similar to the protein expression levels (Fig. 2A). Furthermore, the expression levels of LDLR and miR-33a in HSCs increased along with their activation (Fig. 2B).

image

Figure 2. Down-regulated PPARγ expression and enhanced SREBP2 and miR-33a signaling after HSC activation in vitro. (A) Total and nuclear protein expression (left panel) and mRNA levels (right panel) of PPARγ and SREBP2 in HSCs cultured for 1, 3, 5, or 7 days after isolation from C57BL/6 mice. The relative protein levels are indicated below the corresponding bands. **P < 0.01, compared with the day 1 culture. (B) Quantification of collagen 1α1, collagen 1α2, αSMA, LDLR, and HMGCR mRNA and miR-33a in HSCs cultured for 1, 3, 5, or 7 days after isolation from C57BL/6 mice. Reflecting the activation of HSCs, the mRNA expression levels of collagen 1α1, collagen 1α2, and αSMA gradually increased from day 1 HSCs to day 3 and 5 HSCs to day 7 HSCs. **P < 0.01 and *P < 0.05, compared with the day 1 cultures. (C) Quantification of PPARγ1, SREBP2, and LDLR mRNA and miR-33a (upper panel) and PPARγ protein (lower panel) in quiescent HSCs treated with PPARγ-siRNA. **P < 0.01 and *P < 0.05, compared with the control culture. (D) Quantification of SREBP2 and LDLR mRNA and miR-33a in quiescent HSCs treated with the PPARγ antagonist at the indicated concentrations. **P < 0.01 and *P < 0.05, compared with the control culture. (E) Quantification of PPARγ1, SREBP2, and LDLR mRNA and miR-33a (upper panel), and PPARγ protein (lower panel) in activated HSCs treated with PPARγ1-O/E vector. **P < 0.01, compared with the control culture. (F) Quantification of SREBP2, LDLR, and HMGCR mRNA in activated HSCs treated with SREBP2-siRNA (upper panel). Quantification of LDLR mRNA in quiescent HSCs treated with PPARγ-siRNA and/or SREBP2-siRNA (lower panel). **P < 0.01, compared with the control culture. All data are expressed as means (SEM).

Download figure to PowerPoint

PPARγ-siRNA treatment significantly increased the expression levels of SREBP2, LDLR, and miR-33a in quiescent HSCs (Fig. 2C). Similarly, treatment with the PPARγ antagonist significantly increased the expression levels of SREBP2, LDLR, and miR-33a in quiescent HSCs in a dose-dependent manner (Fig. 2D). On the other hand, overexpression (O/E) of PPARγ1 significantly decreased the levels of SREBP2, LDLR, and miR-33a expression in activated HSCs (Fig. 2E).

SREBP2-siRNA treatment significantly decreased the mRNA expression level of LDLR (Fig. 2F). The addition of PPARγ-siRNA did not affect the mRNA expression level of LDLR in quiescent HSCs treated with SREBP2-siRNA (Fig. 2F).

Enhancement of LDLR Expression and miR-33a Signaling Plays a Role in FC Accumulation in HSCs, Which Subsequently Increases TLR4 Protein Expression Through Suppression of the Endosomal-Lysosomal Degradation Pathway of TLR4

Suppression of LDLR mRNA expression by LDLR-siRNA treatment significantly decreased FC accumulation in HSCs treated with LDL or FBS (Fig. 3A). In HSCs treated with LDL or FBS, FC accumulation significantly decreased with the addition of anti-miR33a and increased with the addition of pre-miR33a (Fig. 3B). Furthermore, FC accumulation in HSCs increased along with their activation (Fig. 3C).

image

Figure 3. FC accumulation in HSCs due to enhanced LDLR expression and miR-33a signaling. (A) Quantification of LDLR mRNA (left panel) and cellular FC and CE (right panel) in HSCs treated with LDLR-siRNA in the presence of LDL or FBS. (B) Quantification of miR-33a (left panel) and cellular FC and CE (right panel) in HSCs treated with anti-miR33a or pre-miR33a in the presence of LDL or FBS. **P < 0.01, compared with the control culture. ##P < 0.01 and #P < 0.05, compared with the FC contents in the corresponding cultures without the addition of LDLR-siRNA, anti-miR33a, or pre-miR33a. (C) Quantification of cellular FC and CE in HSCs cultured for 1, 3, 5, or 7 days after isolation from C57BL/6 mice. **P < 0.01 and *P < 0.05, compared with the day 1 cultures. (D) Quantification of TLR4 mRNA (left panel) and expression of TLR4 protein (right panel) in HSCs cultured for 1, 3, 5, or 7 days after isolation from C57BL/6 mice. The relative protein levels are indicated below the corresponding bands. (E) Expression (left panels) and quantification (right panels) of TLR4 protein in HSCs treated with control-siRNA, LDLR-siRNA, anti-miR33a, pre-miR33a, or control-miR33a in the presence of LDL. **P < 0.01 and *P < 0.05, compared with the control culture. All data are expressed as means (SEM).

Download figure to PowerPoint

TLR4 protein expression, but not mRNA expression, in HSCs increased along with their activation (Fig. 3D). Treatment with LDL significantly increased TLR4 protein expression in HSCs and suppression of LDLR expression significantly decreased it (Fig. 3E). Similarly, the LDL-induced increase in TLR4 protein expression was significantly suppressed by the addition of anti-miR33a and significantly enhanced by the addition of pre-miR33a (Fig. 3E).

Furthermore, treatment with LDL significantly suppressed the ligand-mediated enhanced degradation of TLR4 in HSCs (Fig. 4A). Both chloroquine, an inhibitor of the endosomal-lysosomal pathways, and MG-132, an inhibitor of the proteosomal pathways, significantly increased TLR4 protein expression in HSCs (Fig. 4B). The addition of LDL did not affect the protein expression levels of TLR4 in HSCs treated with chloroquine, whereas it significantly increased the protein levels of TLR4 in HSCs treated with MG-132 (Fig. 4C,D).

image

Figure 4. FC accumulation in HSCs enhances TLR4 protein expression by suppressing the endosomal-lysosomal degradation pathway of TLR4. (A) Expression and quantification of TLR4 protein expression in vehicle-treated or LDL-treated HSCs, 60 minutes after addition of LPS (100 ng/mL), compared with cells not treated with LPS. **P < 0.01 and *P < 0.05, compared with the control culture, without LPS treatment. (B) Expression and quantification of TLR4 protein expression in HSCs treated with MG-132 and/or chloroquine. **P < 0.01 and *P < 0.05, compared with the control culture. (C) Expression and quantification of TLR4 protein expression in HSCs treated with LDL in the presence/absence of MG-132. **P < 0.01 and *P < 0.05, compared with the control culture. (D) Expression and quantification of TLR4 protein expression in HSCs treated with LDL in the presence/absence of chloroquine. **P < 0.01, compared with the control culture. All data are expressed as means (SEM).

Download figure to PowerPoint

FC Accumulation in HSCs Sensitizes These Cells to TGFβ-Induced Activation Through Enhancement of TLR4-Mediated Down-Regulation of Bambi

The mRNA level of Bambi significantly decreased with LPS treatment, and furthermore, the addition of LDL significantly enhanced the decrease in wild-type HSCs (Fig. 5B). A deficiency in TLR4 signaling reversed these decreases (Fig. 5B).

image

Figure 5. FC accumulation in HSCs sensitizes HSCs to TGFβ-induced activation through enhancement of TLR4-mediated down-regulation of Bambi. (A) Quantification of cellular FC and CE in wild-type or TLR4-deficient HSCs, treated or untreated with LDL. **P < 0.01 and ##P < 0.01, compared with the corresponding control culture. (B) Quantification of Bambi mRNA in wild-type or TLR4-deficient HSCs, treated with LPS and/or LDL. **P < 0.01 and *P < 0.05, compared with the corresponding control culture. (C) Quantification of collagen 1α1 and collagen 1α2 mRNA in wild-type or TLR4-deficient HSCs, treated with LPS, TGFβ, and/or LDL. **P < 0.01, compared with the corresponding control culture. (D) Quantification of Bambi mRNA in wild-type or TLR4-deficient HSCs, treated with LDLR-siRNA, control-siRNA, anti-miR33a, or pre-miR33a in the presence of LPS and/or LDL. **P < 0.01, compared with the corresponding control culture. (E) Quantification of collagen 1α1 and collagen 1α2 mRNA in wild-type or TLR4-deficient HSCs, treated with LDLR-siRNA, control-siRNA, anti-miR33a, or pre-miR33a in the presence of LPS, TGFβ, and/or LDL. **P < 0.01 and *P < 0.05, compared with the corresponding control culture. (F) Quantification of Bambi mRNA in HSCs treated with Bambi-siRNA or control-siRNA (left panel). Quantification of collagen 1α1 and collagen 1α2 mRNA in wild-type HSCs, treated with Bambi-siRNA or control-siRNA in the presence of LPS, TGFβ, and/or LDL (right panel). **P < 0.01, compared with the corresponding control culture. All data are expressed as means (SEM).

Download figure to PowerPoint

Wild-type HSCs, pretreated with LPS, demonstrated significant enhancement of collagen 1α1 and 1α2 mRNA expressions when stimulated with TGFβ, and showed a further increase in mRNA expression of collagen 1α1 and 1α2 when treated with LDL (Fig. 5C). A deficiency in TLR4 signaling, however, eliminated these increases (Fig. 5C).

Bambi mRNA expression did not decrease in HSCs treated with LDL, LDLR-siRNA, anti-miR33a, or pre-miR33a in the absence of LPS, but it significantly decreased when HSCs were treated with LPS (Fig. 5D). This decrease was significantly enhanced in cells treated with LDL, whereas treatment with LDLR-siRNA reversed the LDL-induced decrease in Bambi mRNA expression (Fig. 5D). Similarly, treatment with anti-miR33a reversed the LDL-induced decrease in Bambi mRNA expression. On the other hand, treatment with pre-miR33a enhanced the LDL-induced decrease in Bambi mRNA expression (Fig. 5D). These results were in accordance with the results of FC accumulation and TLR4 protein expression in HSCs, and a deficiency in TLR4 signaling reversed all these changes (Fig. 5D).

Treatment with LDLR-siRNA reversed the LDL-induced increase in the mRNA expressions of collagen 1α1 and 1α2 in wild-type HSCs treated with LPS and TGFβ (Fig. 5E). In accordance with the results of FC accumulation and Bambi mRNA expression in HSCs, treatment with anti-miR33a reversed the LDL-induced increase in collagen 1α1 and 1α2 mRNA expression and treatment with pre-miR33a enhanced it (Fig. 5E). As is the case in Bambi mRNA expression, a deficiency in TLR4 signaling canceled all these LDL-induced changes in collagen 1α1 and 1α2 mRNA expression (Fig. 5E). In addition, treatment with Bambi-siRNA reversed the LDL-induced increase in the mRNA expression of collagen 1α1 and 1α2 in HSCs treated with LPS and TGFβ (Fig. 5F). Furthermore, in the same way as in the in vitro study, treatment with antagomirs against miR33a significantly alleviated the activation of HSCs in the mouse model of liver fibrosis induced by carbon tetrachloride (CCl4). This occurred through the suppression of FC accumulation and the subsequent inhibition of TLR4-mediated down-regulation of Bambi in HSCs (Supporting Fig. 8).

Increased Intake of Cholesterol Does Not Impact Liver Fibrosis in NASH in TLR4-Deficient Mice

We used TLR4-deficient mice to assess whether the exacerbation of liver fibrosis in NASH by increased cholesterol intake was dependent on TLR4 signal transduction. Significant differences were not observed in the extent of liver fibrosis or in the hepatic mRNA levels of collagen 1α1, collagen 1α2, and αSMA, between MCD diet-fed and MCD+HC diet-fed TLR4-deficient mice (Fig. 6A-C). Similarly, the increased cholesterol intake did not enhance liver fibrosis in the HF diet-induced NASH in TLR4-deficient mice (Fig. 6D-F).

image

Figure 6. Increased cholesterol intake does not impact liver fibrosis in NASH in TLR4-deficient mice. TLR4-deficient mice (7-8 weeks old; n = 4-7/group) were fed (A-C) the control, HC, MCD, or MCD+HC diet for 8 weeks or (D-F) the control, HC, HF, or HF+HC diet for 20 weeks. (A,D) Hematoxylin and eosin-stained, Masson's trichrome-stained, and αSMA-immunostained sections of representative liver samples. (B,E) Quantification of Masson's trichrome staining (upper panel) and αSMA immunostaining (lower panel). (C,F) Quantification of hepatic collagen 1α1, collagen 1α2, αSMA, and TGFβ mRNA. **P < 0.01, compared with the control diet group. All data are expressed as means (SEM).

Download figure to PowerPoint

SREBP2-Mediated Feedback Regulation of Cholesterol Homeostasis Is Disrupted in HSCs and HSC Activation Further Enhances the Disruption

Nuclear accumulation of hepatic SREBP2 decreased in the two mouse models of NASH and further declined following supplementation with cholesterol (Supporting Fig. 9A). Cholesterol supplementation significantly decreased the hepatic mRNA levels of LDLR and HMGCR, which are downstream molecules of SREBP2, in both the animal models (Supporting Fig. 9B,C).

We next detailed the SREBP2-mediated feedback system of cholesterol homeostasis in hepatocytes and HSCs in vitro. The nuclear form of SREBP2 in hepatocytes was dramatically decreased by treatments with LDL (Fig. 7A) and 25-hydroxycholesterol, which promotes Scap-Insig complex formation.[11] These treatments also significantly decreased the nuclear form of SREBP2 in quiescent HSCs but did not affect that in activated HSCs (Fig. 7A). Quantitative analysis showed that the decrease was significantly enhanced in hepatocytes, compared with HSCs, and quiescent HSCs, compared with activated HSCs (Fig. 7A).

image

Figure 7. The sterol regulatory systems in HSCs are disrupted and dependent on the relative amounts of Scap and Insigs. (A) Expression and quantification of the nuclear form of SREBP2 protein in hepatocytes, quiescent HSCs (qHSCs; cultured for 1 day after isolation), and activated HSCs (aHSCs; cultured for 7 days after isolation) after treatment with LDL, 25-hydroxycholesterol (25-HC), or MβCD/cholesterol complex. **P < 0.01 and *P < 0.05, compared with the corresponding control culture. (B) Expression and quantification of Insig-1, Insig-2, and Scap protein in hepatocytes, qHSCs, and aHSCs. **P < 0.01, compared with the levels in hepatocytes. (C) Expression and quantification of the nuclear form of SREBP2 protein (upper panel) in qHSCs treated with Scap-siRNA, Insig-1-siRNA, or control-siRNA in the presence/absence of LDL, (middle panel) in qHSCs treated with Insig-2-O/E vector or control vector in the presence/absence of LDL, (lower panel) in aHSCs treated with Insig-1-O/E vector, Insig-2-O/E vector, or control vector in the presence/absence of LDL. **P < 0.01 and *P < 0.05, compared with the control culture. (D) Immunoprecipitation analysis of Scap-Insig-1 and Scap-Insig-2 complexes in qHSCs treated with control-siRNA, Scap-siRNA, Insig-1-siRNA, control vector, or Insig-2-O/E vector in the presence of LDL. (E) Immunoprecipitation analysis of Scap-Insig-1 and Scap-Insig-2 complexes in aHSCs treated with control vector, Insig-1-O/E vector, or Insig-2-O/E vector in the presence of LDL. All data are expressed as means (SEM).

Download figure to PowerPoint

MβCD reportedly delivers cholesterol to cells without passing through lysosomes.[12] Treatment with a cholesterol-MβCD complex also dramatically decreased the nuclear form of SREBP2 in hepatocytes (Fig. 7A). This treatment significantly decreased the nuclear form of SREBP2 in quiescent HSCs but did not affect that in activated HSCs (Fig. 7A). Quantitative analysis showed that the decrease was significantly enhanced in hepatocytes, compared with HSCs, and in quiescent HSCs, compared with activated HSCs (Fig. 7A). Scap expression levels were much higher in quiescent and activated HSCs than in hepatocytes (Fig. 7B). However, the Insig-1 expression level in hepatocytes was comparable to that in quiescent HSCs; we did not detect any expression of Insig-1 in activated HSCs (Fig. 7B). Hepatocytes expressed Insig-2 protein, whereas we could not observe any expression of Insig-2 in HSCs (Fig. 7B).

A Scap trypsin cleavage assay[13] was subsequently performed to examine whether or not cholesterol-induced Scap conformational changes occurred in these cells. Scap, without cholesterol-induced conformational changes, yields a protected band of 27 kDa on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), whereas Scap, with the conformational change, yields a protected band of 26 kDa. Our data showed that the cholesterol-induced Scap conformational change in activated HSCs occurred to the same degree as that in quiescent HSCs or hepatocytes (Supporting Fig. 10A,B).

LDL treatment decreased the nuclear level of SREBP2 in quiescent HSCs. Treatment with Scap-siRNA or Insig-2-overexpression vector enhanced the effect, whereas treatment with Insig-1-siRNA counteracted the effect (Fig. 7C, upper and middle). However, LDL treatment did not affect the nuclear level of SREBP2 in activated HSCs; overexpression of Insig-1 or Insig-2 in HSCs significantly decreased the nuclear level of SREBP2 after the addition of LDL (Fig. 7C, lower).

LDL treatment increased the level of the Scap-Insig-1 complex in quiescent HSCs, whereas cotreatment with Scap-siRNA or Insig-1-siRNA reversed this change (Fig. 7D). We could not detect any Scap-Insig-2 complex in quiescent HSCs after the addition of LDL. Overexpression of Insig-2 increased the level of the Scap-Insig-2 complex in LDL-treated quiescent HSCs (Fig. 7D). On the other hand, neither the Scap-Insig-1 nor the Scap-Insig-2 complex could be detected in activated HSCs treated with LDL or not (Fig. 7E). Overexpression of Insig-1 increased the level of the Scap-Insig-1 complex in activated HSCs treated with LDL, and similarly, overexpression of Insig-2 increased the level of the Scap-Insig-2 complex after treatment with LDL (Fig. 7E).

In addition, the feedback regulation system of cholesterol homeostasis impacted the sensitization of HSCs to TGFβ-induced activation, in a manner similar to the FC accumulation system mediated by LDLR or miR33a (Supporting Fig. 11).

HSC Activation in NASH Down-Regulates Insig-1 Expression Through the Suppression of PPARγ Signal Transduction

The Insig-1 expression level was significantly lower in HSCs from the MCD and HF diet-fed groups than in those from the corresponding control diet-fed groups (Fig. 8A,B; Supporting Fig. 12A,B). These decreases were significantly enhanced by the increased intake of cholesterol (Fig. 8A,B; Supporting Fig. 12A,B). We could not detect any difference in the Scap expression level in HSCs among the groups (Fig. 8A,B; Supporting Fig. 12A,B).

image

Figure 8. Down-regulation of Insig-1 expression by HSC activation through the suppression of PPARγ signal transduction. C57BL/6 mice (9 weeks old, male; n = 6-9/group) were fed (A) the control, HC, MCD, or MCD+HC diet for 12 weeks or (B) the control, HC, HF, or HF+HC diet for 24 weeks. (A,B) Expression and quantification of Insig-1 and Scap protein in HSCs isolated from the mice in each group. **P < 0.01 and *P < 0.05, compared with the control diet group. (C) Quantification of Insig-1 mRNA in quiescent HSCs treated with the PPARγ antagonist. **P < 0.01, compared with the control culture. All data are expressed as means (SEM). (D) Schematic of the characteristic mechanisms of FC accumulation in HSCs during the development of liver fibrosis in NASH. FC loading of HSCs is not sufficient to induce activation but serves to enhance activation initiated by TGFβ. Enhanced FC accumulation in HSCs plays an important role in the progression of liver fibrosis in NASH by promoting TLR4 signal transduction through suppression of the endosomal-lysosomal degradation pathway of TLR4, down-regulating the Bambi expression level, and subsequently sensitizing HSCs to TGFβ-induced activation. HSCs are sensitive to FC accumulation because of the high intracellular Scap-to-Insig expression ratio, and furthermore, HSC activation dysregulates their cholesterol metabolism, resulting in further FC accumulation and exaggerating liver fibrosis in a vicious cycle.

Download figure to PowerPoint

Furthermore, Insig-1 protein was abundant in quiescent HSCs but its level declined at days 3 and 5, and day 7 HSCs (Supporting Fig. 12C). We could not detect any significant difference in the Scap expression level among the groups (Supporting Fig. 12C). Similar results were obtained in terms of the mRNA expression levels of Insig-1 and Scap (Supporting Fig. 12C). Treatment with the PPARγ antagonist significantly decreased the Insig-1 expression level in quiescent HSCs in a dose-dependent manner (Fig. 8C).

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

This study showed that increased cholesterol intake accelerated liver fibrosis in the two mouse models of NASH without affecting the degree of hepatocellular injury or Kupffer cell activation. The exacerbation of liver fibrosis mainly involved FC accumulation in HSCs, which increased TLR4 protein levels through suppression of the endosomal-lysosomal degradation pathway of TLR4, down-regulated the expression of the TGFβ pseudoreceptor Bambi, and thereby sensitized the cells to TGFβ-induced activation. This study also showed that FC loading of HSCs is not sufficient to induce activation but serves to enhance activation initiated by TGFβ. These results are compatible with our previous finding[3] that showed that FC accumulation in HSCs increased membrane TLR4 levels; suppressed the HSC expression of Bambi, the TLR4 target gene[14]; and subsequently exaggerated liver fibrosis in mouse models of liver fibrosis.

This study also helped to elucidate the main mechanisms by which HSCs are sensitive to FC accumulation. The SREBP2-mediated feedback system, which plays a major role in maintaining cellular cholesterol homeostasis,[5, 6] was disrupted in HSCs; this disruption could be attributed to high expression of Scap and no expression of Insig-2 in these cells. This could explain why the HC diet significantly reduced SREBP2 signaling in hepatocytes but not in HSCs, and resulted in enhanced FC accumulation in HSCs.

Furthermore, HSC activation sensitized these cells to FC accumulation. Repression of PPARγ signaling underlies HSC transdifferentiation.[15] In the present study, the level of PPARγ decreased along with the activation of HSCs. The suppression of PPARγ signaling in activated HSCs decreased the cellular expression of Insig-1, which resulted in enhancing the disruption of the SREBP2-mediated cholesterol-feedback system. This could partly explain why SREBP2 signaling in HSCs was enhanced, along with their activation, although FC accumulation continued to increase.

In addition, the decreased PPARγ signaling in activated HSCs also enhanced SREBP2 expression and signaling, resulting in enhanced expression of the LDLR, the SREBP2 target gene, in HSCs. As SREBF2 is a bifunctional locus encoding SREBP2 and miR-33a,[10] suppression of PPARγ signaling also increased the level of miR-33a in HSCs, in turn suppressing the levels of NPC1 and ABCA1 (data not shown), which are negatively regulated by miR-33a.[10] These results showed that HSC activation enhanced FC accumulation, in part because of the increased LDLR level and the decreased NPC1 and ABCA1 levels.

The present results suggest that these characteristic mechanisms in HSCs could sensitize the cells to enhanced FC accumulation after increased intake of cholesterol and/or activation of HSCs. They also suggest that such accumulation could play an important role as a mediator of the vicious cycle of HSC activation in NASH (Fig. 8D).

There are two major pathways for cell surface receptor degradation after ubiquitination: a ubiquitin-proteasome pathway and a lysosomal degradation pathway.[16] Our present results showed that FC accumulation in HSCs inhibited the degradation of TLR4, mainly by down-regulating a lysosomal degradation pathway, which resulted in increased levels of TLR4 protein. These results are compatible with our previous report[3] showing that FC accumulation in HSCs could be involved in endosomal-lysosomal dysfunction.

The MCD diet-induced mouse model is commonly used as a model of NASH, and the resulting characteristic pathology of steatosis, mixed cell inflammatory infiltrate, hepatocellular necrosis, and pericellular fibrosis mimics that found in humans with NASH.[17, 18] Nevertheless, the mice do not develop the accompanying metabolic syndrome that is often associated with human NASH. Therefore, we also used an HF diet-induced model of NASH to examine the precise role of cholesterol in the pathophysiology of NASH. As the results were similar in both mouse models of NASH, our findings may indicate a role for cholesterol in the pathophysiology of NASH.

Mari et al.[19] reported that mitochondrial FC loading accounted for hepatocellular sensitivity to TNFα. Furthermore, they showed that the mitochondrial FC content in mouse hepatocytes increased transiently only during the first 6 days of HC feeding, and thereafter returned to its prior level.[19] Our results also showed that chronic HC feeding did not significantly increase mitochondrial FC accumulation in hepatocytes. This could be one reason why an increased intake of cholesterol did not impact the hepatocellular damage in our two mouse models of NASH.

A recent report showed that accumulation of cholesterol in the lysosomes of Kupffer cells increased hepatic inflammation in the mouse model of NAFLD.[20] 27-Hydroxycholesterol is enzymatically generated from mitochondrial cholesterol by the mitochondrial P450 enzyme, Cyp27a1.[21] Further, it mobilizes cholesterol from the lysosomes to the cytoplasm, resulting in a reduction in the accumulation of lysosomal cholesterol in Kupffer cells.[20] In both mouse models of NASH, an increased intake of cholesterol did not affect the lysosomal cholesterol levels in Kupffer cells, nor did it impact the mitochondrial cholesterol levels or Cyp27a1 expression levels in Kupffer cells. These could be some reasons why increased cholesterol intake did not accelerate Kupffer cell activation in our mouse models of NASH.

In conclusion, FC accumulation in HSCs was enhanced mainly by two mechanisms: enhancement of both SREBP2 and miR-33a signaling through the suppression of PPARγ signaling along with HSC activation and disruption of the SREBP2-mediated cholesterol-feedback system in HSCs, which was characterized by a high Scap-to-Insig ratio and exaggerated by the down-regulation of Insig-1 through the suppression of PPARγ signaling along with HSC activation. Enhanced FC accumulation in HSCs plays an important role in the progression of liver fibrosis in NASH by promoting TLR4 signal transduction through suppression of the endosomal-lysosomal degradation pathway of TLR4, and subsequently sensitizing HSCs to TGFβ-induced activation. HSC activation dysregulates their cholesterol metabolism, resulting in further FC accumulation and exaggerating liver fibrosis in a vicious cycle (Fig. 8D). We believe that the characteristic mechanisms of FC accumulation in HSCs should be further studied as potential targets to treat liver fibrosis in liver diseases including NASH.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

The authors thank Mina Kitazume and Miho Takabe (Keio University) for helpful advice and technical assistance, and Drs. Ikuo Inoue and Makoto Seo (Saitama Medical School) for helpful discussion and critical comments.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
hep26604-sup-0001-suppfig1.tif7986KSupporting Information Figure 1.
hep26604-sup-0002-suppfig2.tif2549KSupporting Information Figure 2.
hep26604-sup-0003-suppfig3.tif141KSupporting Information Figure 3.
hep26604-sup-0004-suppfig4.tif3432KSupporting Information Figure 4.
hep26604-sup-0005-suppfig5.tif3449KSupporting Information Figure 5.
hep26604-sup-0006-suppfig6.tif5771KSupporting Information Figure 6.
hep26604-sup-0007-suppfig7.tif252KSupporting Information Figure 7.
hep26604-sup-0008-suppfig8.tif189KSupporting Information Figure 8.
hep26604-sup-0009-suppfig9.tif275KSupporting Information Figure 9.
hep26604-sup-0010-suppfig10.tif223KSupporting Information Figure 10.
hep26604-sup-0011-suppfig11.tif231KSupporting Information Figure 11.
hep26604-sup-0012-suppfig12.tif143KSupporting Information Figure 12.
hep26604-sup-0013-suppinfo.doc165KSupporting Information
hep26604-sup-0014-supptable1.doc55KSupporting Information Table 1.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.