SEARCH

SEARCH BY CITATION

Keywords:

  • electron microscopy;
  • chronic hepatitis;
  • CREB3L3;
  • eIF-2-alpha;
  • unfolded protein response;
  • physiology;
  • pathology;
  • inflammation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

In hepatocytes, the accumulation of unfolded proteins in the endoplasmic reticulum (ER) causes ER stress and the unfolded protein response (UPR), mediated by the ER-resident stress sensors ATF-6, IRE1, and PERK. UPR-responsive genes are involved in the fate of ER-stressed cells. Cells carrying hepatitis C virus (HCV) subgenomic replicons exhibit in vitro ER stress and suggest that HCV inhibits the UPR. Since in vivo ER homeostasis is unknown in livers with chronic HCV infection, we investigated ER stress and the UPR in liver samples from untreated patients with chronic hepatitis C (CHC), in comparison with normal livers. Electron microscopy, western blotting, and real-time RT-PCR were used in liver biopsy specimens. Electron microscopy identified features showing ER stress in hepatocyte samples from patients with CHC; however, ‘ER-stressed’ hepatocytes were found in clusters (3-5 cells) that were scattered in the liver parenchyma. Western blot analysis confirmed the existence of hepatic ER stress by showing activation of the three ER stress sensors ATF-6, IRE1, and PERK in CHC. Real-time RT-PCR showed no significant induction of UPR-responsive genes in CHC. In contrast, genes involved in the control of diffuse processes such as liver proliferation, inflammation, and apoptosis were significantly induced in CHC. In conclusion, livers from patients with untreated CHC exhibit in vivo hepatocyte ER stress and activation of the three UPR sensors without apparent induction of UPR-responsive genes. This lack of gene induction may be explained by the inhibiting action of HCV per se (as suggested by in vitro studies) and/or by our finding of the localized nature of hepatocyte ER stress. Copyright © 2010 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

In cells, newly synthesized secretory and membrane-associated proteins are correctly folded and assembled in the endoplasmic reticulum (ER) 1–3. Once ER homeostasis is perturbed by various pathological conditions, newly synthesized unfolded proteins accumulate in the ER, resulting in ER stress 1–3. To cope with accumulated unfolded ER proteins, mammalian cells trigger a specific adaptive response called the unfolded protein response (UPR) 1–3 (Figure 1). There are three distinct signalling pathways that are induced by ER stress mediated by three ER-resident stress sensors, ie the activating transcription factor 6 (ATF-6), the inositol requiring enzyme 1 (IRE1), and the double-stranded RNA-activated protein kinase-like ER kinase (PERK) 4–6.

thumbnail image

Figure 1. Adaptative signalling of the unfolded protein response (UPR). The three UPR transducers PERK, ATF-6, and IRE1 are associated with BiP in their monomeric inactive form. Upon ER stress-dependent accumulation of unfolded proteins in the ER lumen, Bip is released from these sensors, inducing their homodimerization and subsequent activation. Activated PERK phosphorylates the translation initiation factor eIF-2α, which attenuates the general translation rate and induces the translation of selective mRNAs with inhibitory uORFs in their 5′ UTR, such as the transcriptional factor ATF-4. GADD34, a target of ATF-4 activation, negatively regulates the eIF-2α-mediated inhibition of translation. The release of BiP from ATF-6 induces its trafficking into the Golgi apparatus, where it is cleaved by proteases S1P/S2P and thereby enters the nucleus to activate the transcription of XBP1 and other UPR target genes. Activated IRE1 catalyses the splicing of XBP1 mRNA to allow the translation of mature XBP-1 protein. XBP-1 mediates transcriptional up-regulation of numerous ER stress-dependent genes. The downstream effectors of these three signalling pathways combinatorially induce the expression of genes involved in the ER protein folding capacity, such as chaperones. In parallel, the ER-associated degradation (ERAD) process is activated to remove misfolded proteins from the ER lumen in coordination with the proteasome machinery

Download figure to PowerPoint

ATF-6 is synthesized as a precursor protein and anchored to the ER membrane, where it is retained by the HSP70-class ER chaperone BiP [also known as glucose-related protein (GRP) 78] 1–3. In response to ER stress, ATF-6 is released from BiP and transported to the Golgi complex, where ATF-6 undergoes regulated intramembrane proteolysis, ie sequential cleavage by two proteases, S1P and S2P 2. The processed form of ATF-6, which may target genes since it is a transcription factor [belonging to the basic-region leucine zipper (bZIP) family 7], translocates to the nucleus 1, 5–7. The second UPR branch involves IRE1 and X-box-binding protein 1 (XBP-1) 1–4. IRE1 contains both serine/threonine kinase and ribonuclease domains. Under normal conditions, XBP1 mRNA is translated, but its product is a weak transcriptional activator with a short protein half-life 2. During ER stress, activated IRE1 cuts 26 nucleotides from XBP1 mRNA to generate spliced XBP1 mRNA, which encodes the more stable and transcriptionally active XBP-1S protein (which belongs to the family of bZIP transcription factors 2). The third UPR branch is mediated by PERK, which is a serine/threonine protein kinase that phosphorylates the alpha subunit of eukaryotic translation initiation factor 2 (eIF2-alpha) 1–3. Phosphorylation of eIF2-alpha subsequently inhibits global protein synthesis 1–3. Paradoxically, eIF2-alpha phosphorylation induces translation of activating transcription factor 4 (ATF4) mRNA into the bZIP transcription factor ATF-4 [2].

The UPR activates the transcription of certain genes (eg those encoding chaperones, lectins, and calcium pump) that serve to increase the ER's protein folding capacity as needed. UPR signalling can protect cells from ER stress by expanding the amount of ER in the cell, enhancing the degradation of misfolded proteins (eg via XBP-1S induction of EDEM1), and reducing the synthesis of new proteins (via eIF2-alpha phosphorylation) 1–3. In addition, the UPR may also promote an adaptive response, known as the ER overload response (EOR), which involves IRE1-mediated activation of nuclear factor (NF)-kappaB, a transcription factor involved in anti-apoptotic responses 8. The EOR has also been shown to involve GSK3-beta, which inhibits p53-induced apoptosis 9. However, when the adaptive responses are not sufficient to relieve the ER stress, apoptotic pathway(s) are activated via IRE1 or eIF2-alpha/ATF-4 pathways 1, 2. IRE1 is associated with the apoptosis-regulating BCL-2 protein family members BAK and BAX 10. IRE1 signalling has been shown to activate the c-Jun NH2-terminal kinase (JNK) and/or caspase-12 1, 2. ER stress-induced apoptosis may also result from the induction of the ATF-4 target gene DDIT3 [which codes for C/EBP homologous protein-10 (CHOP), a bZIP transcription factor 1, 2, 7].

Chronic hepatitis C (CHC) is a leading cause of cirrhosis and hepatocellular carcinoma worldwide 11. HCV is an enveloped flavivirus with a 9.6-kb single-strand RNA genome 11. This genome serves as a template for replication and as a viral messenger RNA for production of the virus. HCV enzymes (ie NS2–3 and NS3–4A proteases, NS3 helicase, and NS5B RdRp) are essential for HCV replication 11. The development of a subgenomic HCV RNA replicon capable of replication in the human hepatoma cell line Huh7 has been a significant advance 12, 13. Recently, complete replication of HCV in cell culture has been achieved 14.

HCV core synthesis, processing, and folding occur in the ER. Envelope proteins E1 and E2 reside in the ER lumen and the viral replicase is assumed to localize on ER-derived membranes 15–17, leading to the possibility of inducing the ER-resident stress sensors. In vitro studies have shown that ER stress may be triggered by HCV subgenomic replicons 18, 19, structural or non-structural proteins 20–23. In vitro studies also suggest that HCV may interfere with the UPR to inhibit it 24. Together, in vitro studies suggest that HCV may trigger ER stress and inhibit the resulting UPR. However, experimental models for HCV infection have limitations. For instance, these models remained unable to release infectious HCV particles 12, 13. Therefore, in vivo studies are of major importance. To date, there is no information on in vivo ER homeostasis and the UPR in CHC. Thus, this study investigated in vivo ER stress and the UPR in liver samples from patients with CHC.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Patients and tissue specimens

Percutaneous liver biopsies from 28 untreated patients with CHC were studied (Table 1); 13 had mild fibrosis and 15 advanced fibrosis (Supporting information, Supplementary materials and methods). Percutaneous liver-biopsy specimens were selected from a cohort of adult patients with untreated CHC followed at Beaujon Hospital (Clichy, France). All CHC patients had antibodies against HCV (AxSYM® Anti-HCV, Abbott) and detectable serum HCV RNA (TMA, Bayer's Versant® HCV RNA Qualitative Assay). HCV genotyping was performed (sequencing) and serum HCV RNA was quantified [Bayer's Versant® HCV RNA 3.0 Assay (bDNA)] for all patients. No patient had clinical evidence of hepatic decompensation. The following conditions were excluded: positive HBsAg, human immunodeficiency virus infection, autoimmune hepatitis, haemochromatosis, alpha1-antitrypsin deficiency, and Wilson's disease. Characteristics of these patients are presented in Table 1. Pretreatment liver biopsies from patients before decision to treat was made, both immediately frozen liver tissue (stored at − 80 °C) and fixed paraffin-embedded tissue (for histology), were available.

Table 1. Characteristics of 28 patients with chronic hepatitis C
VariablePatients
No28
Gender (male, n)19
Age (years)45.3 ± 8.2
Source of infection (n) 
 Blood transfusion9
 Intravenous drug use9
 Unknown10
Body mass index (kg/m2) ( ± SD)21.3 ± 1.9
Blood glucose (mmol/l) ( ± SD)5.1 ± 1.2
ALT (IU/l, median)114 ± 43
HCV genotype (n) 
 115
 24
 37
 42
Viral load (mean, log10 IU/ml)5.2 ± 0.4
 F113
Stage (n) 
F3–F415

Ten patients had mild chronic hepatitis B. All of these patients had positive hepatitis B surface antigen (HBsAg), detectable serum HBV DNA [Bayer's Versant® HBV DNA 3.0 Assay (bDNA)], and were untreated. No patient had clinical evidence of hepatic decompensation. Moreover, in the present study, the following conditions were excluded: positive antibodies against HCV, human immunodeficiency virus infection, autoimmune hepatitis, haemochromatosis, alpha1-antitrypsin deficiency, and Wilson's disease.

Liver biopsies from seven adults with normal liver histological aspects were studied 25.

Percutaneous liver-biopsy specimens were obtained from adults with mild elevated alanine aminotransferase activity, with no cause of liver disease (medication, alcohol, chronic viral hepatitis, autoimmune processes, and metabolic disease). All of these adults gave their informed consent for the study. All of these liver tissue specimens were histologically normal (ie absence of inflammation, fibrosis, and pathological pattern). The study was approved by the Local Committee of Ethics and conformed to the ethical guidelines of Helsinki. All patients gave their informed consent prior to liver biopsy.

Transmission electron microscopy

Blinded electron microscopy was performed in 13 patients with mild CHC and five histologically normal controls. In fact, since cirrhosis and advanced fibrosis are associated with several features (apoptosis, necrosis, cholestasis, etc), we hypothesized that ER stress activation might not be specific for HCV but related to the development of cirrhosis. Thus, we decided to study patients with chronic HCV infection and mild liver disease. Liver tissue was immediately fixed by immersion in a 2% solution of glutaraldehyde buffered with 0.2 M cacodylate buffer and post-fixed in osmium tetroxide before embedding in epoxy resin. Ultra-thin sections stained with uranyl acetate and lead citrate were examined with a Jeol 10 10 (Tokyo, Japan) electron microscope.

Real-time RT-PCR

In previous studies using the same technological approach, we have shown that several altered molecular pathways are involved in CHC 26, 27. The method of real-time quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) has been described in detail elsewhere 26, 27 and in the Supporting information, Supplementary materials and methods. By studying the literature, we selected 17 genes that have been shown to be induced by ‘traditional’ ER stressors [under in vivo conditions (in particular, in the liver) or in vitro in hepatoma or any other human cells] 28–31 or by HCV subgenomic replicons, structural or non-structural proteins (under in vitro conditions in hepatoma cells) (Supporting information, Supplementary Table 1). Oligonucleotide primer sequences used are listed in Supporting information, Supplementary Table 2.

Table 2. Unfolded protein response (UPR)-responsive genes in livers from patients with mild or advanced HCV-related fibrosis relative to normal livers
   Advancedp valuep value
 NormalMild HCV-related fibrosisHCV-relatedMild HCV-relatedMild vs advanced
Gene(n = 7)(n = 13)fibrosis (n = 15)fibrosis vs normalHCV-related fibrosis
  1. Values are medians with range. HCV = hepatitis C virus. For other abbreviations, see Supporting information, Supplementary Table 1. The Ntarget values of the samples were subsequently normalized such that the mean of the normal histological liver Ntarget values was 1.

XBP11.17 (0.65–1.31)0.87 (0.22–1.39)0.58 (0.16–1.14)0.190.19
Spliced XBP11.19 (0.33–1.45)1.29 (0.27–1.98)0.79 (0.26–2.31)0.750.31
Ratio of spliced XBP1 to XBP11.02 (0.42–1.25)1.25 (0.82–2.00)1.36 (0.81–2.03)0.020.56
ATF61.00 (0.67–1.18)1.09 (0.71–3.96)1.23 (0.47–2.71)0.150.91
CREBL11.00 (0.86–1.26)0.85 (1.06–2.90)0.79 (0.41–1.96)0.160.60
ATF41.00 (0.80–1.24)0.90 (0.67–2.61)1.10 (0.56–2.19)0.430.20
HSPA50.96 (0.71–2.21)0.83 (0.26–2.94)0.84 (0.26–2.40)0.510.91
HSP90B11.05 (0.56–1.55)0.87 (0.32–1.43)0.83 (0.037–1.67)0.550.71
CANX1.20 (0.70–1.54)0.77 (0.14–2.26)0.62 (0.27–1.40)0.190.62
CALR1.20 (0.58–1.86)1.02 (0.28–2.45)1.01 (0.57–2.02)0.690.98
EDEM11.04 (0.56–1.51)0.76 (0.28–1.66)0.52 (0.33–1.28)0.380.13
ATP2A21.00 (0.44–2.45)0.91 (0.36–1.48)0.78 (0.51–3.48)0.360.88
PPP1R15A1.03 (0.49–1.56)0.81 (0.44–1.54)1.21 (0.29–1.86)0.040.19
DDIT31.17 (0.00–1.95)1.19 (0.06–2.47)0.88 (0.31–2.72)0.580.43
NFE2L21.00 (0.44–2.45)1.19 (0.66–1.78)1.24 (0.42–2.54)0.200.66

Western blot analysis

The protein expression levels of EDEM, ATF-4, BiP/GRP78, eIF2-alpha, phosphorylated eIF2-alpha, and beta-actin were measured in liver samples. ATF-6 alpha, ATF-6 beta, XBP-1, PERK, eIF2-alpha, EDEM, ATF-4, BiP/GRP78, and beta-actin (Santa Cruz Biotechnology, Inc, Santa Cruz, CA, USA), and antibodies against phospho-PERK (Thr980) and phospho-eIF2-alpha (Ser51) (Cell Signaling Technology, Beverly, MA, USA) were measured.

Statistical analysis

Values (medians with range) were tested using the Kruskall–Wallis test. Differences between three groups were judged significant at confidence levels greater than 95% (p < 0.05). The relative mRNA levels shown in Table 2 (calculated as described in the Supporting information, Supplementary materials and methods) show the abundance of the target relative to the endogenous control (TBP), in order to normalize the starting amount and quality of total RNA. Similar results were obtained with a second endogenous control, RPLP0 (also known as 36B4) (data not shown).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Electron microscopy shows altered hepatocyte ER morphology in livers from patients with mild HCV-related fibrosis

Electron microscopy was used to compare livers from patients with mild CHC with normal livers. The appearance of the ER in hepatocytes from normal livers was normal, with cisternae of rough ER regularly organized into stacks around the nucleus, while smooth ER appeared as small vesicles or tubules on the periphery of the stacks. Cisternae lumens were narrow and on the external face, many groups of bound ribosomes were visible (Figure 2A). In contrast, in livers from most patients with mild CHC (9/13), both parts of the ER (smooth and rough) were dilated and disorganized (Figure 2B). However, many bound ribosomes were still visible on the surface of the rough ER (Figure 2B). These alterations differ from those known to occur in necrotic hepatocytes, which also have dilated ER but are associated with a significant loss of ribosomes. Together, these morphological findings are similar to those found elsewhere in ER-stressed pancreatic endocrine cells 32. It should be noted that livers with ER-stressed hepatocytes exhibited changes that were not diffuse but clustered into small groups of 3–5 cells, with no preferential lobular localization (Figure 2C). In addition, the proportion of cells with ER changes differed from one patient to another (range 10–30%), indicating inter-individual variability. Interestingly, there was no significant correlation between viral load and the proportion of cells with ER changes. Finally, the other structures of the hepatocytes such as the nucleus and the mitochondria were normal in all patients. Moreover, there were no significant changes in ‘non-hepatocyte’ cells.

thumbnail image

Figure 2. Electron microscopy shows altered hepatocyte ER morphology in liver biopsy specimens from patients with mild HCV-related liver fibrosis. (A) The appearance of endoplasmic reticulum (ER) in hepatocytes from normal livers was normal, with cisternae of the rough ER (RER) regularly organized in stacks around the nucleus. Lumens of the cisternae were narrow and on the external face, many groups of bound ribosomes were visible (arrows). (B) In contrast, in hepatocytes from livers with mild hepatitis C, the two parts of the ER (rough and smooth, RER and SER) were dilated and disorganized. However, at the surface of the rough ER, bound ribosomes were clearly visible, with maintenance of many ribosomes (arrows). Fine granular deposits in the lumen of the RER could represent unfolded protein (stars). N = nucleus; M = mitochondria. (C) At this low magnification, electron microscopy shows a cluster of three hepatocytes (H) where the ER (arrows) is dilated

Download figure to PowerPoint

Livers from patients with HCV-related fibrosis exhibit activation of the three proximal ER-resident stress sensors

Next we investigated whether ER stress was associated with induction of UPR in livers from patients with mild or advanced HCV-related fibrosis compared with normal livers.

ATF-6 pathway

There are two ATF-6 isoforms, alpha and beta (Supporting information, Supplementary Table 1). In ER-stressed cells, ATF-6 is released from the ER membrane and is converted (by cleavage due to regulated intra-membrane proteolysis) from a 90 kD protein (p90ATF-6, alpha or beta) to a 50 kD protein (p50ATF-6, alpha or beta) which translocates to the nucleus 1–3. Thus, we investigated whether ER stress induces hepatic p50ATF-6 alpha and/or p50ATF-6 beta expression in livers with mild HCV-related fibrosis. For this, we measured the expression level of the two ATF-6 isoforms using specific antibodies. In normal livers, there was no expression of the two isoforms, while in mild HCV-related fibrosis, the expression levels of the two p50ATF-6 isoforms were significantly increased (Figure 3), a finding consistent with ER stress-induced ATF-6 activation. Hepatic expression levels of p50ATF-6 alpha and p50ATF-6 beta were significantly higher in advanced fibrosis compared with mild fibrosis (Figure 3).

thumbnail image

Figure 3. Activation of ER-resident stress sensors in livers from patients with chronic hepatitis C. (A) Western blot analyses with anti-p50ATF-6α, anti-p50ATF-6β, anti-XBP-1, anti-PERK, anti-PERK phospho-specific Thr980 (P-PERK), and anti-housekeeping β-actin were performed on cell homogenates from normal livers (n = 2), and livers from patients with mild (n = 4) or advanced (n = 3) hepatitis C virus (HCV)-related fibrosis. Of note, the anti-XBP-1 antibody recognizes both 29 kD XBP-1 (encoded by ‘unspliced’ XBP1 mRNA) and 50 kD XBP-1S (encoded by ‘spliced’ XBP1 mRNA). (B) Bars represent the means ( ± SD) of ‘normalized’ densitometric values for proteins shown in the left panel. For p50ATF-6 isoforms, the ratio of p50ATF-6 (α or β) to β-actin is given. The ratios of P-PERK to total PERK and of XBP-1S to XBP-1 are given for P-PERK and XBP-1S, respectively. Both p50ATF-6 isoforms were expressed in livers with mild fibrosis but not in normal livers. The expression levels of the p50ATF-6α and p50ATF-6β isoforms were significantly higher in livers with advanced fibrosis than in mild fibrosis. XBP-1S protein was expressed in advanced fibrosis but not in mild fibrosis or normal livers. It should be noted that the expression levels of XBP-1 were lower in livers with HCV-related fibrosis than in normal livers. PERK was phosphorylated on Thr980 in mild fibrosis but not in normal livers. PERK phosphorylation was significantly more marked in advanced than in mild fibrosis

Download figure to PowerPoint

IRE1 pathway

ER stress activates IRE1 1–3. Activated IRE1 cuts XBP1 mRNA into spliced XBP1 mRNA, encoding transcriptionally active XBP-1S protein 1–3. Thus, the relative hepatic mRNA expression of spliced XBP1 was examined and the ratio of spliced XBP1 mRNA to XBP1 mRNA, which indirectly reflects IRE1 endoribonuclease activity 34, was calculated. The ratio of spliced XBP1 to XBP1 mRNAs in mild CHC was significantly increased compared with normal livers (Table 2), suggesting ER stress-elicited activation of IRE1. However, there was no significant increase in the protein expression levels of XBP-1S in mild fibrosis (Figure 3). Although the ratio of spliced XBP1 to XBP1 mRNAs was similarly increased in livers with advanced fibrosis and in those with mild fibrosis (Table 2), significant increases in XBP-1S protein expression levels were found in the former but not the latter (Figure 3). There is no clear explanation for these findings in advanced fibrosis.

PERK pathway

ER stress activates the eIF2-alpha kinase PERK 1–3. PERK activation is associated with its phosphorylation on Thr980 2 and results in eIF2-alpha phosphorylation on Ser21. Thus, we studied the hepatic expression levels of total PERK, total eIF2-alpha, phosphorylated PERK (Thr980), and phosphorylated eIF2-alpha (Ser51) by western blot with specific antibodies. PERK was phosphorylated in mild fibrosis but not in normal livers, while the expression levels of total PERK protein were similar in both groups (Figure 3). Consistent with PERK activation in mild fibrosis, eIF2-alpha was strongly phosphorylated in diseased livers (Figure 4), a finding not observed in normal livers (Figure 4). EIF2-alpha phosphorylation results in the attenuation of translation of most mRNAs 29. However, eIF2-alpha phosphorylation elicits ‘paradoxical’ translation of ATF4 mRNA into ATF-4 1, 2. By using a specific antibody, we found that ATF-4 protein was expressed in mild fibrosis but not in normal liver (Figure 4). The levels of PERK phosphorylation on Thr980 were significantly higher in advanced fibrosis than in mild fibrosis (Figure 3), indicating further PERK activation with fibrosis progression. However, unlike livers with mild fibrosis, livers with advanced fibrosis did not have any phosphorylation of eIF2-alpha (Figure 4). Consistent with this, ATF-4 protein levels were significantly lower in advanced fibrosis than in mild fibrosis (Figure 4).

thumbnail image

Figure 4. Altered expression of EDEM, ATF-4, phosphorylated eIF2-α in livers from patients with chronic hepatitis C according to the severity of liver fibrosis. (A) Western blot analyses using anti-EDEM, anti-ATF-4, anti-BiP/GRP78, anti-eIF2-α, anti-eIF2-α phospho-specific Ser51 (P-eIF2-α), and anti-housekeeping gene (β-actin) antibodies were performed on liver cell homogenates from patients with mild (n = 4) or advanced (n = 4) hepatitis C virus (HCV)-related fibrosis. Cell homogeneates from patients with normal livers (n = 3) were also investigated. (B) Bars represent the means ( ± SD) of ‘normalized’ densitometric values for proteins shown in the left panel. The expression of EDEM, ATF-4, and BiP/GRP78 was normalized using corresponding β-actin. The expression of P-eIF2-α was normalized using corresponding values of total eIF2-α. The protein expression levels of EDEM and ATF-4 were significantly lower and the levels of eIF2-α phosphorylation significantly higher in advanced fibrosis than in mild fibrosis. There was also a trend in the decrease in the expression level of BiP/GRP78

Download figure to PowerPoint

In summary, CHC is associated with activation of the three hepatic UPR sensors and this activation increases with fibrosis progression.

No induction of hepatic UPR-responsive genes in HCV-related fibrosis

The UPR has been shown to result in the induction genes encoding bZIP transcription factors [XBP1, ATF6 (encoding ATF-6 alpha), CREBL1 (encoding ATF-6 beta), ATF4], chaperones (HSPA5, HSP90B1), lectins (CANX, CALR, EDEM1) and calcium pump (ATP2A2) 5, 6, PPP1R15A (encoding GADD34, a protein which triggers a negative feedback mechanism resulting in eIF2-alpha dephosphorylation 1–3), and DDIT3 (encoding the pro-apoptotic protein CHOP). The UPR may also be associated with the induction of NFE2L2 [encoding nuclear factor erythroid 2-related factor 2 (Nrf2), a Cap'n'Collar bZIP transcription factor that dimerizes with ATF-4 to activate the antioxidant response element 30]. Thus, the relative levels for all of these mRNAs were examined in mild or advanced CHC compared with normal livers. We found no significant induction of any of these mRNAs in CHC compared with normal livers (Table 2). PPP1R15A mRNA expression was even down-regulated in diseased livers compared with normal livers (Table 2). In addition, western blot analysis showed an accumulation of BiP/GRP78 (the protein encoded by HSPA5) and EDEM1 in mild fibrosis but not in normal liver (Figure 4). Together, these findings suggest that post-transcriptional mechanisms favour the accumulation of BiP/GRP78 and EDEM1 proteins in mild CHC. Interestingly, the protein expression levels of EDEM1 were significantly lower in livers with advanced fibrosis compared with mild fibrosis (Figure 4).

Induction of genes involved in liver proliferation, inflammation, and apoptosis in HCV-related fibrosis

We then asked whether the expression of genes known to be involved in diffuse processes such as hepatocyte proliferation, liver inflammation, and apoptosis 11–26 was increased in diseased livers. Thus, we measured the expression mRNA levels of MKi67 (proliferation), JUN (a result of JNK activation involved in proliferation and inflammation), CREB3L3 mRNA (also known as CREBH, encoding CREB-H 31, a hepatocyte-specific bZIP transcription factor of the ATF subfamily which serves as a link between inflammation and the acute-phase response), NF-kappaB-induced anti-apoptotic mRNAs (cFLAR, GADD45B, BCL2A1, and IRE3), the short variant of IER3 called IER3S, and p53-inducible pro-apoptotic mRNAs (BBC3, PMAIP1, BAX, and FAS).

In livers with mild CHC, there were significant increases in the relative mRNA levels of MKi67, JUN, CREB3L3, BCL2A1, and IER3S (encoding the pro-apoptotic immediate early response 3 isoform short variant) (Table 3). Compared with mild CHC, advanced CHC had significant increases in the mRNA levels of MKi67, JUN, IER3S, BAX, and FAS (Table 3). In contrast, cFLAR expression was down-regulated (Table 3). Thus, our results confirm that real-time RT-PCR is able to detect changes in gene expression related to diffuse intra-hepatic processes.

Table 3. Expression of genes involved in the control of proliferation, inflammation, and apoptosis in livers from patients with mild or advanced HCV-related fibrosis relative to normal livers
  Mild HCV-relatedAdvanced HCV-relatedp valuep value
 NormalfibrosisfibrosisMild HCV-relatedMild vs advanced
Gene(n = 7)(n = 13)(n = 15)fibrosis vs normalHCV-related fibrosis
  1. Values are medians with range. HCV = hepatitis C virus. For other abbreviations, see Supporting information, Supplementary Table 1. The Ntarget values of the samples were subsequently normalized such that the mean of the normal histological liver Ntarget values was 1.

MKI670.99 (0.69–2.20)2.02 (0.13–3.81)3.78 (1.63–24.39)0.04< 0.01
JUN0.99 (0.69–1.40)1.63 (0.78–3.46)3.17 (1.61–19.16)0.02< 0.01
CREB3L31.00 (0.73–1.49)1.64 (0.80–2.80)1.80 (0.79–3.72)0.020.79
CFLAR0.99 (0.29–1.33)0.96 (0.27–1.98)0.43 (0.03–0.98)0.860.01
GADD45B1.15 (0.62–3.71)0.37 (0.11–3.15)0.50 (0.23–1.58)0.080.71
BCL2A11.00 (0.72–2.07)2.35 (0.90–10.36)4.71 (0.78–34.35)< 0.010.20
IER30.99 (0.18–2.42)2.15 (0.48–33.32)5.09 (1.83–22.22)0.070.06
IER3S1.00 (0.82–1.46)1.99 (0.86–5.12)4.93 (1.84–14.14)< 0.01< 0.01
BBC30.99 (0.49–1.89)1.56 (0.37–4.00)2.49 (0.95–7.16)0.110.23
PMAIP11.00 (0.27–2.94)0.68 (0.17–3.75)0.99 (0.14–5.63)0.630.79
BAX1.01 (0.71–1.54)0.83 (0.29–2.16)1.86 (0.83–2.81)0.900.02
FAS1.00 (0.77–2.16)1.82 (0.54–3.53)3.19 (1.60–5.83)0.07< 0.01

Hepatic mRNA expression associated with mild liver fibrosis may differ between HBV and HCV infections

To determine whether the genes that are deregulated during mild chronic HCV-induced fibrosis (the transition from normal to mild fibrosis) are specific for HCV infection, the mRNA expression levels of selected genes were measured in livers with mild HBV fibrosis.

UPR-responsive genes

Alterations observed in patients with HCV infection [ie increased ratio of spliced XBP1 mRNA to XBP1 mRNA (indicating engagement of the IRE1 pathway) and down-regulated PPP1R15A (see above and Table 4)] were not found in livers with mild HBV fibrosis (Table 4). In contrast, certain genes whose expression was unchanged in mild HCV fibrosis were up-regulated in mild HBV fibrosis, including ATF6, ATF4, and NFE2L2 (Table 4).

Table 4. Gene expression in livers from patients with HBV mild fibrosis, relative to normal livers
 NormalMild HBV-relatedp value
Genes(n = 7)fibrosis (n = 10)HBV vs normal
  • Values are medians with range. HBV = hepatitis B virus. For other abbreviations, see Supporting information, Supplementary Table 1. The Ntarget values of the samples were subsequently normalized such that the mean of the normal histological liver Ntarget values was 1.

  • *

    Other genes are those which are involved in the control of proliferation, inflammation, and apoptosis.

UPR-responsive genes   
 XBP11.17 (0.65–1.31)0.67 (0.21–1.38)0.04
 Spliced XBP11.19 (0.33–1.45)0.51 (0.20–2.00)0.22
 Ratio of spliced XBP1 to XBP11.02 (0.42–1.25)1.10 (0.49–1.62)0.55
 ATF61.00 (0.67–1.18)1.45 (1.09–2.82)< 0.01
 CREBL11.00 (0.86–1.26)1.01 (0.63–1.62)0.96
 ATF41.00 (0.80–1.24)1.37 (1.00–3.29)0.02
 EDEM11.04 (0.56–1.51)0.93 (0.40–1.83)0.78
 HSPA50.96 (0.71–2.21)0.99 (0.04–4.73)0.45
 HSP90B11.05 (0.56–1.55)0.88 (0.54–1.49)0.75
 CANX1.20 (0.70–1.54)0.84 (0.33–1.76)0.22
 CALR1.20 (0.58–1.86)0.73 (0.27–2.65)0.25
 ATP2A21.00 (0.44–2.45)0.82 (0.05–5.52)0.45
 PPP1R15A1.03 (0.49–1.56)0.80 (0.15–1.26)0.12
 DDIT31.17 (0.00–1.95)1.09 (0.11–1.67)0.84
 NFE2L21.00 (0.44–2.45)1.32 (0.85–2.88)0.02
Other genes*   
 MKI670.99 (0.69–2.20)2.33 (0.46–4.98)0.05
 JUN0.99 (0.69–1.40)1.63 (0.71–3.20)0.02
 CREB3L31.00 (0.73–1.49)1.17 (0.81–5.84)0.17
 CFLAR0.99 (0.29–1.33)0.50 (0.30–1.10)0.07
 BCL2A11.00 (0.72–2.07)4.51 (0.46–14.59)0.02
 IER30.99 (0.18–2.42)2.02 (0.24–11.22)0.05
 IER3S1.00 (0.82–1.46)1.77 (0.80–12.02)0.02
 GADD45B1.15 (0.62–3.71)0.62 (0.28–3.10)0.20
 BBC30.99 (0.49–1.89)1.07 (0.32–3.70)0.62
 PMAIP11.00 (0.27–2.94)2.45 (0.19–5.04)0.28
 BAX1.01 (0.71–1.54)0.95 (0.02–2.17)0.72
 FAS1.00 (0.77–2.16)1.96 (0.98–6.04)0.03
Genes involved in liver proliferation, inflammation, and apoptosis

CREB3L3, which was up-regulated in mild HCV infection, was not altered in patients with mild HBV-related fibrosis. FAS, whose expression was unchanged in the early stage of chronic HCV infection, was up-regulated in patients with mild HBV-related fibrosis (Table 4). BCL2A1, IER3S, and JUN were up-regulated in both groups of patients. Finally, GADD45B, BBC3, PMAIP1, BAX, and cFLAR were not induced in either group of patients.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

In this study, hepatocytes in livers from untreated patients with mild CHC had dilated and disorganized rough ER on electron microscopy. Experimental studies have shown that dilation and disorganization of rough ER occur when misfolded or unfolded proteins accumulate in the ER lumen, ie when there is ER stress 1–3. Thus, our morphological findings show that there is hepatocyte ER stress in mild CHC. Moreover, we found evidence of the activation of the three ER-resident stress sensors, ATF-6, IRE1, and PERK, indicating proximal engagement of UPR pathways in these livers. In addition, there was a more marked engagement of the three ER-resident stress sensors in advanced CHC, suggesting an increase in ER stress with fibrosis progression. Our results are supported by in vitro studies suggesting that HCV induces perturbation of ER homeostasis. HCV structural proteins E1 and E2 may be disulphide-linked and form misfolded aggregates in the ER lumen 21. HepG2 cells expressing the HCV core develop ER calcium depletion 20, a mechanism of ER stress 2. There is indirect evidence of engagement of the UPR (ie ER stress) in cells expressing HCV subgenomic replicons 18, 19, structural proteins (ie core 20, E2 alone 21, E1 alone or combined with E2 22), or the non-structural protein NS4B 23.

It should be noted that in our study morphological evidence of ER stress was found only in some clusters of hepatocytes. Interestingly, studies have shown that HCV infects only a limited number of hepatocytes 33, 34. Recently, by combining 2-photon microscopy with virus-specific, fluorescent, semiconductor quantum dot probes, it has been shown that HCV involves a limited number of hepatocytes 33. Furthermore, the low average levels of HCV-RNA in biopsy samples might be explained by focal distribution of infected hepatocytes 34. Thus, ER stress may exist only in a small number of hepatocytes because of a ‘localized’ pattern of HCV infection. Future studies are needed to correlate ER stress to the presence of HCV.

This study shows that although there was ER stress and activation of the three UPR sensors in livers with CHC, full induction of UPR-responsive genes was not found in these livers. This finding may have two explanations which are not mutually exclusive. First, in vitro studies suggest that HCV per se may inhibit the UPR 24. HCV subgenomic replicons activate ATF-6, induce HSP5 mRNA coding for BiP but inhibit its translation 18. HCV subgenomic replicons were found to induce XBP1 mRNA splicing and also to inhibit XBP-1S trans-activating activity, EDEM induction, and degradation of unfolded proteins 19. E2 (but not E1) protein induces the HSP5 mRNA coding for BiP but not the protein itself 21. In addition, E2 proteins may inhibit the activity of PERK 35 or the interferon-induced, double-stranded RNA-activated protein kinase (PKR) 36. Moreover, there is evidence that initiation of protein synthesis by HCV is refractory to eIF2-alpha phosphorylation 37.

Our morphological findings may provide another explanation for the lack of induction of UPR-responsive genes in CHC liver-biopsy specimens. Indeed, the lack of gene induction may be related to the fact that only hepatocytes in small clusters were ‘ER-stressed’ and surrounded by numerous cells that did not experience ER stress (see above). Instead, processes such as proliferation, inflammation, and apoptosis are known to be diffuse in livers with CHC 11, 26, 38–40. Inflammation is a diffuse feature in CHC driven by the persistent expression of chemokines 26, 38, 40. It is also accepted that apoptosis of hepatocytes is a prominent feature of CHC 39. Consistent with this, we found that genes involved in the control of proliferation, inflammation, and apoptosis were significantly induced in livers from patients with CHC.

Livers from mice with dietary-induced or genetic (ob/ob) obesity exhibit chronic ER stress which may result in hepatic insulin resistance 41. Thus, in patients with CHC, excess weight, but not HCV per se, may trigger hepatic ER stress. However, this hypothesis is unlikely since our patients had a normal body mass index and normal blood glucose concentrations.

Finally, this study investigated gene expression profiles in livers with mild HBV-related fibrosis. It is important to note that the expression gene profile differed between HBV and HCV infection. Five genes including ATF6, ATF4, NFE2L2, IER3, and FAS were up-regulated in livers with mild HBV-related fibrosis but not in those with mild HCV-related fibrosis. Interestingly, although ATF6 and ATF4 are known to be induced by XBP-1S, no evidence of IRE1 activation or induction of other XBP-1S target mRNAs (eg CREBL1 or EDEM1) was found. Moreover, it should be noted that ATF6, ATF4, and NFE2L2 each encode a protein that belongs to the family of bZIP transcription factors 7. Further studies are needed to investigate the ‘propensity’ of HBV infection to induce genes coding for bZIP transcription factors.

In conclusion, livers from patients with CHC exhibit in vivo hepatocyte ER stress and activation of the three ER-resident UPR sensors without apparent induction of UPR-responsive genes. This lack of gene induction may be explained by the inhibiting action of HCV per se and/or by our finding of the localized nature of hepatocyte ER stress.

SUPPORTING INFORMATION ON THE INTERNET

The following supporting information may be found in the online version of this article.

Supplementary materials and methods. List of genes investigated and oligonucleotide primer sequences used.

Table S1. List of genes investigated in patients with HCV-related fibrosis.

Table S2. Oligonucleotide primer sequences used.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank Alain Grodet for his technical assistance in electron microscopy and Jean-Pierre Lagneau for medical illustrations. This study was supported by the INSERM, LFB and the association pour la Recherche sur le cancer (ARC). CG received a Dame Sheila Sherlock Fellowship from EASL. RM is in receipt of an Interface INSERM-AP-HP Fellowship.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Note: References 42–44 are cited in the Supporting information to this article.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
path_2703_sm_suppinfoS1.doc60KSupporting Information: Table S1. List of genes investigated in patients with HCV-related fibrosis.
path_2703_sm_suppinfoS2.doc38KSupporting Information: Table S2. Oligonucleotide primer sequences used.
path_2703_sm_suppinfoS3.doc33KSupporting Information: Supplementary materials and methods.

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.