Cleavage of Mitochondrial Antiviral Signaling Protein
in the Liver of Patients with Chronic Hepatitis C
Correlates with a Reduced Activation of the
Endogenous Interferon System
Pantxika Bellecave,1* Magdalena Sarasin-Filipowicz,2* Olivier Donz´ e,3Audrey Kennel,1J´ erˆ ome Gouttenoire,1
Etienne Meylan,4,5Luigi Terracciano,6J¨ urg Tschopp,4Christoph Sarrazin,7Thomas Berg,8
Darius Moradpour,1† and Markus H. Heim2†
some but not all patients with chronic hepatitis C (CHC). Patients with a pre-activated IFN
system are less likely to respond to the current standard therapy with pegylated IFN-?. Mito-
chondrial antiviral signaling protein (MAVS) is an important adaptor molecule in a signal
transduction pathway that senses viral infections and transcriptionally activates IFN-?. The
analyzed liver biopsies from 129 patients with CHC to investigate whether MAVS is cleaved in
vivo and whether cleavage prevents the induction of the endogenous IFN system. Cleavage of
MAVS was detected in 62 of the 129 samples (48%) and was more extensive in patients with a
high HCV viral load. MAVS was cleaved by all HCV genotypes (GTs), but more efficiently by
GTs 2 and 3 than by GTs 1 and 4. The IFN-induced Janus kinase (Jak)-signal transducer and
activator of transcription protein (STAT) pathway was less frequently activated in patients with
expression level of the IFN-stimulated genes IFI44L, Viperin, IFI27, USP18, and STAT1. We
CHC is in part regulated by cleavage of MAVS. (HEPATOLOGY 2010;51:1127-1136.)
pathways that lead to type I interferon (IFN) (IFN-? and
nfection with the hepatitis C virus (HCV) leads to
chronic hepatitis C (CHC) in 50% to 80% of indi-
viduals. The recognition of HCV by the host triggers
IFN-?) production and to the induction of an antiviral
state.1,2To establish persistent infection, HCV has
evolved numerous strategies to evade and counteract the
Abbreviations: ?CT, difference in the cycle threshold; cEVR, complete early virological response; CHC, chronic hepatitis C; EVR, early virological response; FL,
full-length; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GT, genotype; HCV, hepatitis C virus; IFN, interferon; ISG, interferon-stimulated gene; Jak, Janus
kinase; mAb, monoclonal antibody; MAVS, mitochondrial antiviral signaling protein; mRNA, messenger RNA; NR, nonresponse; PNR, primary nonresponse; RIG-I,
retinoic acid-inducible gene I; STAT, signal transducer and activator of transcription protein; VL, viral load.
From the1Division of Gastroenterology and Hepatology, Centre Hospitalier Universitaire Vaudois, University of Lausanne, Lausanne, Switzerland; the2Department
of Biomedicine and Division of Gastroenterology and Hepatology, University Hospital Basel, Basel, Switzerland;3Apotech Corporation, Epalinges; the4Department of
the6Pathology Institute, University Hospital Basel, Basel, Switzerland; the7Department of Medicine I, J. W. Goethe University Hospital, Frankfurt, Germany;
8Medizinische Klinik m. S. Hepatologie und Gastroenterologie, Charite ´ Campus Virchow-Klinikum, Berlin, Germany.
*These authors have equally contributed to this work
†These authors have equally contributed to this work.
Received July 21, 2009; accepted November 4, 2009.
Address reprint requests to: Markus H. Heim, M.D., Department of Biomedicine, University Hospital Basel, Hebelstrasse 20, CH-4031 Basel, Switzerland. E-mail:
email@example.com; Fax: (41)-61-265-53-52 or Darius Moradpour, M.D., Division of Gastroenterology and Hepatology, Centre Hospitaliser Universitaire
Vaudois, Rue du Bugnon 44, CH-1011 Lausanne, Switzerland. E-mail: Darius.Moradpour@chuv.ch; Fax (41)-21-314-47-18.
Copyright © 2009 by the American Association for the Study of Liver Diseases.
Published online in Wiley InterScience (www.interscience.wiley.com).
Supported by the Swiss National Science Foundation (grants 3100A0-122447 and 32-116106), the Swiss Cancer League/Oncosuisse (grants OCS-01762-08-2005
and KLS-01832-02-2006) and the De ´sire ´e and Niels Yde Foundation.
Potential conflict of interest: Nothing to report.
Additional Supporting Information may be found in the online version of this article.
tified the HCV NS3-4A serine protease as a key viral
protein blocking innate immune pathways. NS3-4A
cleaves and thereby inactivates the caspase recruitment
domain–containing essential adaptor protein mitochon-
drial antiviral signaling protein (MAVS)7(also known as
caspase recruitment domain adaptor inducing IFN-?,8
interferon-? promoter stimulator protein 1,9and virus-
induced signaling adaptor10) in the retinoic acid-induc-
ible gene-I (RIG-I) viral RNA-sensing pathway.8MAVS
is located at the outer mitochondrial membrane and as-
sociates with RIG-I through its caspase recruitment do-
main. This interaction, on activation of a kinase complex
comprising TRAF family member-associated NF-?B ac-
tivator (TANK)-binding kinase 1 (TBK1) and inhibitor
of the cytoplasmic IFN regulatory factor 3 and IFN reg-
ulatory factor 7, which dimerize and translocate to the
nucleus, where they induce the transcription of IFN-?
within an almost canonical NS3-4A cleavage site and re-
sults in dislocation of the protein from the outer mitochon-
drial membrane.8,16HCV NS3-4A also targets TIR-
domain-containing adapter-inducing interferon-? (TRIF),
a key adaptor molecule in the Toll-like receptor 3 (TLR3)
double-strand RNA-sensing pathway.16Hence, HCV may
lular proteins essential for the induction of the first-line im-
Despite its ability to inactivate key components of the
viral sensory pathways, HCV triggers an ongoing IFN
response during chronic infections in chimpanzees1and
level of IFN-stimulated gene (ISG) expression among pa-
tients with CHC. Moreover, activation of the endoge-
nous IFN system is linked to the response to the current
standard treatment with pegylated IFN-? and ribavirin.
imens taken before therapy are poor responders to treat-
a favorable response to therapy.2,17,18
Interference of HCV with the innate immune re-
sponse, by cleaving MAVS or TRIF, could explain the
variability of ISG pre-activation in CHC patients. There
is evidence from biochemical analyses and from cell cul-
ture experiments that HCV triggers IFN-? expression
through the RIG-I pathway,19and, as outlined previ-
with the RIG-I pathway by NS3-4A–mediated cleavage
of MAVS.8,16Cleavage of MAVS has been reported in
the current study, we (1) validated and extended the ob-
servations on MAVS cleavage in a large panel of well-
characterized liver biopsy specimens from patients
infected with different HCV genotypes (GTs), (2) deter-
mined whether the extent of MAVS cleavage correlates
with activation of the endogenous IFN system in vivo,
and (3) correlated differences in cleavage and inactivation
of this crucial adaptor molecule with treatment response,
HCV viremia, and GT as well as histological grading and
staging. Our results support a role of MAVS cleavage in
the HCV-mediated control of antiviral responses in vivo
in the liver of patients with CHC.
Materials and Methods
Antibodies. An anti-MAVS rabbit polyclonal anti-
serum and mouse monoclonal antibodies (mAbs) were
raised against an Escherichia coli–expressed recombinant
protein representing amino acids 160 through 450 of
MAVS. The immunoglobulin G2b mAb designated as
Switzerland) under the designation Adri-1 and AT107,
respectively. MAb AC-15 against beta-actin was from
Sigma (St. Louis, MO).
Cell Lines. Huh-7.5 cells21and a subgenomic HCV
replicon that served as controls for detection of MAVS in
its full-length (FL) and cleaved forms were provided by
Charles M. Rice (The Rockefeller University, New York,
Liver Biopsies and Patient Data. Liver biopsy spec-
imens from patients with CHC (n ? 150) and controls
tic workup. Grading and staging of CHC was performed
according to the Metavir classification. A specimen was
obtained for histopathological examination and the pa-
with local ethical committees. Serum HCV RNA was
quantified using the COBAS AmpliPrep/COBAS Taq-
man HCV-Test and the Cobas Amplicor Monitor from
Roche Molecular Systems. Patient characteristics are
shown in Table 1.
Immunoblot Analyses. Proteins were extracted by ho-
mogenization of biopsy samples in a lysis buffer containing
50 mM Tris. Cl pH 8.0, 150 mM NaCl, 1% NP40, 0.5%
deoxycholate, 0.1% sodium dodecyl sulfate, 1 mM sodium
orthovanadate, 10 mM NaF, and a cocktail of protease in-
hibitors (Complete Protease Inhibitor, Roche Diagnostics,
as previously described,22using horseradish peroxidase–
1128 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.HEPATOLOGY, April 2010
ern Blotting Detection Kit (Amersham, Du ¨bendorf,
Switzerland). Densitometric scanning was performed with
an ImageScanner (Amersham), and the bands correspond-
well as beta-actin were quantified with the ImageQuant TL
Immunohistological Staining. Standard indirect im-
munoperoxidase procedures were used for immunohisto-
chemistry (ABC-Elite, Vectra Laboratories, Burlingame,
CA). Four-mm-thick sections were cut from paraffin
blocks, rehydrated, pretreated for 20 minutes in ER2 so-
lution, incubated with a rabbit mAb against phosphory-
lated STAT1 (p-STAT1) (Cell Signaling, Bioconcept,
Allschwil, Switzerland), and counterstained with hema-
toxylin. The entire procedure was performed with an au-
tomated stainer from Vision BioSystems (Newcastle
upon Tyne, UK). For quantification of the p-STAT1
staining, samples were divided into four categories ac-
cording to the proportion of positive hepatocyte nuclei
(?: ?5%; ?: 5%-33%; ??: 34%-66%; ???: ?66%
of hepatocytes with positive nuclear staining).
Measurement of ISG Messenger RNA Levels. RNA
extracted from human liver tissue was used for quantifi-
cation of STAT1, IP10, USP18, IFI27, Viperin and
IFI44L messenger RNAs (mRNAs). Total RNA was ex-
tracted using the RNeasy Mini Kit (Qiagen, Basel, Swit-
zerland) according to manufacturer’s instructions. RNA
was aliquoted and stored at ?75°C. RNA was reverse
transcribed by Moloney murine leukemia virus reverse
transcriptase (Promega Biosciences, Wallisellen, Switzer-
land) in the presence of random hexamers (Promega) and
deoxynucleoside triphosphates. The SYBR-PCR reac-
tions were performed using the SYBR green PCR master
TCGACAGTCA-3? and 5?-ACCTTCCCCATGGTGT-
CTGA-3?; STAT1, 5?-TCCCCAGGCCCTTGTTG-3?
and 5?-CAAGCTGCTGAAGTTGGTACCA-3?; IP10,
GGTACAGCGTACGGTTCT-3?; USP18, 5?-CTCAGT-
CCCGACGTGGAACT-3? and 5?-ATCTCTCAAGCGC-
CATGCA-3?; IFI27, 5?-CCTCGGGCAGCCTTGTG-3?
and 5?-AATCCGGAGAGTCCAGTTGCT-3?; Viperin,
5?-CTTTGTGCTGCCCCTTGAG-3? and 5?-TCCAT-
GGCTGCAGAT-3? and 5?-CTCTCTCAATTGCACC-
AGTTTCC-3?. The difference in the cycle threshold
(?Ct) value was derived by subtracting the Ct value for
GAPDH, which served as internal control, from the Ct
value for transcripts of interest. All reactions were run in
duplicate, using an Applied Biosystems Prism 7000 Se-
quence Detection System. Messenger RNA expression
levels were calculated relative to GAPDH from the ?Ct
values, using the formula 2??Ct.
Table 1. Patient Characteristics
*Mean age of patients with CHC was 44.5 years (SD ? 11.6; range ? 18-80 years). Eighty-four were men and 66 women. One hundred forty-six patients were
white, two were Asian, and two were African American. Cardif analysis was performed in 129 patients with CHC and 39 controls. Samples obtained at the University
Hospital Basel (86 CHC; 16 controls) were subjected to immunohistochemical p-STAT1 staining as well as measurements of intrahepatic viral load and ISG mRNAs
†Controls included biopsy specimens from patients with chronic hepatitis B (n ? 13), nonalcoholic (n ? 7) and alcoholic (n ? 1) steatohepatitis, primary biliary
cirrhosis (n ? 6), autoimmune hepatitis (n ? 5), drug-induced liver disease (n ? 5), primary sclerosing cholangitis (n ? 1), and normal liver tissue adjacent to
adenocarcinoma metastasis (n ? 8).
‡GT was not determined in one patient.
§Eight patients are still under treatment or ?6 months after the end of treatment. Treatment was interrupted in four patients.
Patients* CHC (150) Controls (46)†
HCV genotype‡ (N) 1 (77) 2 (16) 3 (47) 4 (9)
Baseline HCV RNA (IU/mL)
3.1 × 106 (std. error 4.4 × 105; range 4.4 × 102 – 3.6 × 107)
METAVIR grade (N) A0 (0) A1 (52) A2 (66) A3 (29)
METAVIR stage (N)* F0-1 (35) F2 (56) F3 (31) F4 (27)
Treatment response (N) § SVR (69) Relapse (17) NR (42)
HEPATOLOGY, Vol. 51, No. 4, 2010BELLECAVE, SARASIN-FILIPOWICZ, ET AL.1129
Quantification of Intrahepatic HCV RNA. One
microgram total RNA isolated from biopsy specimens of
manufacturer’s instructions using a pathogen-specific RT
primer mix (PrimerDesign, Southampton, UK) designed
for the in vitro quantification of all HCV genotypes.
HCV RNA was quantified using a pathogen-specific
primer/probe mix (PrimerDesign) and the Taqman Uni-
versal PCR Master Mix (Applied Biosystems, manufac-
tured by Roche, NJ). Fluorescence was detected through
the FAM channel of the Applied Biosystems 7000 Se-
quence Detection System, and copy number of HCV
RNA per microgram total RNA was calculated according
to the standard curve obtained using control template
Statistical Methods. Correlations were assessed using
the Spearman coefficient. Comparisons between two
groups or between multiple groups were performed with
the Mann-Whitney test and one-way analysis of variance,
respectively, using GraphPad Prism version 4.00 for
Macintosh (GraphPad Software, San Diego, CA). A P ?
0.05 was considered as statistically significant.
MAVS Is Cleaved in Liver Tissue from Patients
with Chronic Hepatitis C. The clinical characteristics
of all CHC patients and controls who were part of this
study are summarized in Table 1. Liver biopsy samples
from 129 patients with CHC and 39 controls were ana-
lyzed by western blot using mAb IID12 (Adri-1) specific
for MAVS. Both forms of MAVS, FL and cleaved, were
detected in liver biopsy specimens from patients with
CHC, whereas only FL MAVS was detected in controls
(Fig. 1A, B). For all analyses, signal intensities of FL and
cleaved MAVS were normalized to ?-actin. The percent-
age of MAVS cleavage (cleaved/total MAVS ? 100) var-
ied widely among the 129 CHC patients, ranging from 0
to 76% (mean, 18%; standard deviation 21%) (Fig. 1B).
Cleavage of MAVS was detected in 62 of the 129 patients
showed a normal distribution ranging from 7% to 76%
(mean, 37%; standard deviation 16%). Only FL MAVS
was detected in the remaining 67 patients.
Because only the FL form of MAVS is functional, and
because it is unknown whether and how fast the cleaved
MAVS product is degraded in cells, we also performed all
our analyses using the FL MAVS/beta-actin signal inten-
sity ratio. The amount of FL MAVS showed strong neg-
ative correlation with the percentage of cleaved MAVS
(Spearman r ?-0.57, P ? 0.0001; data not shown).
MAVS within both the CHC and control patient groups
in CHC compared with control samples (0.48 versus
0.77; P ? 0.0007). Immunoblotting with the polyclonal
MAVS antiserum AT107 yielded results similar to those
conclude that HCV-induced cleavage of MAVS reduces
the amount of the functional FL MAVS in a considerable
proportion of patients with CHC.
Cleavage of MAVS was independent from the Metavir
tigated HCV GTs were able to cleave MAVS. Samples
from patients infected with the easier-to-treat GTs 2 or 3
had significantly lower amounts of FL MAVS compared
with the difficult-to-treat GTs 1 and 4 (P ? 0.009; Fig.
1D). There was no significant difference between GT 2/3
vir activity grade, or fibrosis score. Therefore, the differ-
ence most likely reflects different intrinsic properties of
HCV GTs with respect to MAVS cleavage.
Correlation of Viral Load and Cleavage of MAVS.
Because in vitro studies demonstrated a very efficient
cleavage of MAVS by the HCV NS3-4A protease,8,16,20
we assessed whether patients with a high VL show more
MAVS cleavage in the liver. Indeed, there was a weak but
statistically significant inverse correlation between the
P ? 0.011; Fig. 2A) and a strong positive correlation
between the percentage of cleaved MAVS and serum VL
(Spearman r ? 0.53, P ? 0.0001; Fig. 2B). The correla-
tion remained significant when only the 62 samples with
some detectable degree of MAVS cleavage were included
in the analysis (Fig. 2C). The mean VL was significantly
higher in the group of 62 patients with detectable MAVS
cleavage than in 67 patients without any detectable cleav-
age product in the liver (Fig. 2D). We also measured
intrahepatic HCV VL in a subset of 46 patients, and, in
agreement with published data,23,24there was a very
strong correlation between intrahepatic and serum VL
(Fig. 2E). As a confirmation of the results obtained with
serum VL, the intrahepatic VL correlated negatively with
the amount of functional FL MAVS and positively with
the percentage of cleaved MAVS (Fig. 2F). Taken to-
gether, we provide strong evidence that high viral replica-
tion is correlated with increased cleavage of MAVS and
reduced amounts of the functional FL form of MAVS.
Induction of the Endogenous IFN System in Hepa-
tocytes of a Subset of Patients with CHC. CHC pa-
interferon alpha and ribavirin show an up-regulated IFN
system in the liver before treatment initiation when com-
pared with patients with a complete early virological re-
1130 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.HEPATOLOGY, April 2010
sponse (cEVR).2,17This activation of the endogenous
IFN system is specific to CHC and is not found in pa-
and Wieland and Chisari25). Expression levels of four se-
lected ISG mRNAs (STAT1, IP10, USP18, IFI27) were
high in pretreatment liver biopsy specimens of CHC pa-
drop of VL at week 12 of treatment), as compared with
patients with a cEVR, or with patients with chronic hep-
atitis B and controls (Fig. 3A; Kruskal-Wallis test, P ?
0.0001). The mechanisms responsible for a preactivated
IFN system in the liver seen in a subgroup of HCV pa-
the ISG up-regulation occurs. To elucidate whether the
increase in ISG transcripts, measured using RNA ex-
tracted from a heterogeneous liver biopsy, results from an
activated type I IFN-induced Jak-STAT signaling path-
way specifically in hepatocytes, we assessed levels of
and eight controls (histologically confirmed healthy liver
tified, and each biopsy sample was assigned to one of four
categories according to the number of stained nuclei
icant correlation of nuclear p-STAT1 staining in hepato-
cytes and the mRNA expression of selected ISGs (Fig.
3B). We therefore propose that the elevated ISG levels
observed in livers of patients with CHC originate from
hepatocytes with an IFN-?/IFN-?-induced activation of
the Jak-STAT signal transduction pathway.
MAVS Cleavage Negatively Correlates with the Ac-
tivation Status of the Endogenous IFN System. The
observed interindividual differences in MAVS cleavage in
patients with CHC (Fig. 1A, B) provide an attractive
hypothesis to explain the differences in preactivation of
the endogenous IFN system in the liver of these patients.
Extensive cleavage of MAVS in some patients could pre-
Fig. 1. MAVS is cleaved specifically in patients with chronic hepatitis
C. (A) Lysates from liver biopsies of patients with CHC (lanes 4-7) as well
as a control (lane 3) were separated by 10% sodium dodecyl sulfate
polyacrylamide gel electrophoresis and analyzed by immunoblot using
mAb IID12 (Adri 1) against MAVS or mAb AC-15 against beta-actin.
Arrows indicate full-length (FL) and cleaved MAVS (CL). Naı ¨ve Huh-7.5
cells (lanes 1 and 8) and Huh-7.5 cells harboring a subgenomic HCV
replicon (lanes 2 and 9) served as controls for FL and cleaved MAVS. (B)
MAVS cleavage in patients with CHC and in controls. Western blot signal
for FL and cleaved MAVS were quantified and normalized to beta-actin.
Cleaved MAVS was calculated as percentage of total MAVS. (C) FL MAVS
normalized to beta-actin in patients with CHC and in controls. The
difference between the two patients groups is statistically significant
(P ? 0.0007, Mann-Whitney test). (D) Genotype (GT) 2 and 3 patients
have significantly fewer FL MAVS compared with GT 1 and 4 patients
(P ? 0.009, Mann-Whitney test). N ? number of patients in each group.
HEPATOLOGY, Vol. 51, No. 4, 2010 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.1131
vent the transcriptional induction of IFN-? in HCV-
infected hepatocytes, thereby preventing the autocrine
and paracrine activation of the Jak-STAT pathway and
the up-regulation of ISGs. These patients would not have
would have a strong induction of the endogenous IFN
system. To test this hypothesis, we performed correlation
analyses of MAVS cleavage with nuclear p-STAT1 stain-
ing in hepatocytes or with induction of ISGs in the liver
(Fig. 4). The extent of MAVS cleavage differed signifi-
cantly between the four categories of nuclear p-STAT1
staining (one-way analysis of variance, P ? 0.023, R2?
0.153), with a significant linear trend between the groups
(slope ? ?4.58, R2? 0.149, P ? 0.0025; Fig. 4A). The
percentage of MAVS cleavage also correlated with the
induction of the ISGs IFI44L, Viperin, IFI27, USP18,
and STAT1 (Fig. 4B, C, D, E, and F), but not IP10 (Fig.
4G). FL MAVS showed significant correlation only with
IFI27 mRNA, but not with the other five investigated
ISGs (data not shown). Interestingly, we found more
virin treatment compared with PNR patients (Fig. 4H).
This difference was independent of GT, because it per-
sisted even after stratification of the data according to GT
(Supporting Fig. 1). As outlined, pre-activation of the
endogenous IFN system is a strong predictor of NR to
cleavage between EVR and PNR patients, together with
the correlation of MAVS cleavage with pre-activation,
therefore supports an important role of MAVS-depen-
dent signaling for the induction of the endogenous IFN
system in patients with CHC.
Activation of the Endogenous IFN System and HCV
VL. We next analyzed whether the HCV VL correlates
with the activation status of the endogenous IFN system.
Because high VL positively correlates with MAVS cleav-
pre-activation (Fig. 4), one could predict that patients
with high VL have less pre-activation of the IFN system.
However, knowing that high VL is associated with poor
response to therapy,26and poor response correlates with
pre-activation,2,17patients with high VL should more of-
ten have a pre-activated endogenous IFN system.
Consistent with these conflicting observations, we
found no significant differences in serum or intrahepatic
VL and the amount of nuclear p-STAT1 staining (Fig. 5)
and no correlation of VL with ISG expression (data not
shown). Our data, therefore, support neither a negative
nor a positive correlation of VL with pre-activation of the
endogenous IFN system. Most likely, this is reflecting the
and the viral sensory pathways of the host.
The cleavage and inactivation of MAVS by the HCV
NS3-4A protease provides a conceptual framework that
could explain why many CHC patients do not activate
their endogenous IFN system in the liver. Such a mecha-
Fig. 2. Correlation of HCV viral load (VL) with cleavage of MAVS. (A)
Western blot signals for FL MAVS were normalized to the beta-actin bands
and then correlated with the serum HCV VL of 129 patients with CHC. There
was a weak significant negative correlation (r ? ?0.22, P ? 0.012). (B)
patients with CHC (r ? 0.53, P ? 0.0001). (C) Correlation of serum VL with
the percentage of cleaved MAVS in the 62 of the 129 patients with MAVS
67 patients without any cleavage of MAVS (r ? 0.35, P ? 0.005). (D)
Serum VL is significantly higher in the group of patients with cleaved MAVS
compared with patients with only FL MAVS in the liver biopsy samples (P ?
0.0001, t test). (E) Intrahepatic HCV RNA (measured in a subgroup of 46
patients) strongly correlates with serum VL (r ? 0.71, P ? 0.0001). (F)
Intrahepatic VL positively correlates with the percentage of cleaved MAVS
(r ? 0.58, P ? 0.0001).
1132 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.HEPATOLOGY, April 2010
nism would predict that HCV infections with a very effi-
cient cleavage of MAVS do not (or only weakly) induce
the transcriptional activation of the IFN-? gene, thereby
preventing the autocrine and paracrine positive feedback
loop through the Jak-STAT pathway that typically limits
the replication and spread of many viruses. The work
presented in this paper supports a role of MAVS cleavage
in the HCV-mediated control of antiviral responses in
vivo. However, we also provide evidence that MAVS
cleavage cannot be the only factor affecting the activation
status of the endogenous IFN system in the liver of pa-
tients with CHC.
MAVS cleavage can be detected in almost half of the
patients with CHC and is found in infections with all
HCV GTs tested (Fig. 1A). Cleavage of MAVS is specific
for hepatitis C, because it was never detected in patients
with other chronic liver diseases, including chronic hepa-
3 cleave MAVS more extensively than the difficult-to-treat
GTs 1 and 4 (Fig. 1D). Accordingly, MAVS cleavage was
2 and 3 than with GTs 1 and 4 (56.6% versus 42.6%, data
not shown). Given the role of MAVS in IFN-? induction,
one would predict that GT 2 and 3 infections would less
often induce activation of the endogenous IFN system. In-
deed, we recently reported a lower rate of ISG induction in
pretreatment biopsy specimens of patients infected with
GTs 2 and 3 when compared with GTs 1 and 4.2In agree-
in GT 2 and 3 patients than in GT 1 and 4 patients (Sup-
porting Fig. 2).
HCV GTs 2 and 3 may also be more successful in
establishing a persistent infection, because they more ef-
ficiently cleave MAVS and thereby hamper innate im-
mune responses. However, the limited data that are
spontaneously cleared during the acute phase than infec-
tions with GT 1,27but another study reported higher
response in clearing HCV infections are also reflected by
the fact that many chronically infected patients have a
strong up-regulation of hundreds of ISGs in the liver.2
There is apparently no simple correlation between the
degree of ISG up-regulation and viral elimination in hep-
Fig. 3. Analysis of selected ISGs in liver biopsy specimens from patients with chronic hepatitis C or chronic hepatitis B and from subjects without
chronic liver disease. (A) The mean expression levels of selected ISG mRNAs (STAT1, IP10, USP18, IFI27) in pretreatment liver biopsy specimens
of patients are significantly different in the four study groups: patients with CHC and complete early virological response (cEVR), patients with CHC
and primary nonresponse (PNR; less than 2 log10drop of VL at week 12), patients with chronic hepatitis B (CHB) and controls (Kruskal-Wallis test,
P ? 0.0001). The mean expression of all four ISGs was significantly different between cEVR and PNR patients (Dunn’s Multiple Comparison test,
P ? 0.01) and between PNR patients and controls (Dunn’s Multiple Comparison test, P ? 0.001). There was no significant difference between CHB
patients and controls. N ? number of patients in each group. (B) Quantification of p-STAT1 nuclear staining in hepatocytes by immunohistochemical
analysis of liver biopsy specimens of CHC patients (?: 5% of hepatocytes; ?: 5%-33%; ??: 34%-66%; ??? ? ?66%) and subjects without
liver disease (controls ? C) and correlation to mRNA levels of four selected ISGs (STAT1, IP-10, USP18, IFI27; expression values relative to GAPDH).
There were significant differences of the mean ISG expression between the five groups (shown are the P-values obtained with the Kruskal-Wallis test).
In Dunn’s Multiple Comparison test, significant (P ? 0.01) differences were found between the ??? group (?66% nuclear p-STAT1 staining) and
the control group and between the ??? and the ?group (?5% nuclear p-STAT1 staining) for all four ISGs, and between the ??? group and
the ? group (5%-33% nuclear p-STAT1 staining) for STAT1, IP10, and USP18. N ? number of patients in each group.
HEPATOLOGY, Vol. 51, No. 4, 2010 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.1133
Our model predicts an inverse correlation between
MAVS cleavage and the activation of the endogenous
IFN system. Indeed, we found that the mean percentage
of MAVS cleavage was significantly lower in patients
showing a strong activation of the Jak-STAT pathway, as
4A). Also, the individual expression levels of five classical
ISGs showed inverse correlations with MAVS cleavage
(Fig. 4B-F). IFI44L, Viperin, IFI27, and USP18 were
chosen for the analysis because high expression of these
genes in liver biopsy specimens of CHC patients is pre-
dictive of NR to pegylated IFN-?/ribavirin treat-
endogenous IFN system. We also analyzed STAT1, be-
cause it is the central signal transducer of type I IFN
signaling, and IP-10 because it has been used as a serum
marker for response to pegylated IFN-?/ribavirin,29al-
absence of significant correlation of IP10 with MAVS
cleavage (Fig. 4G) may be attributable to the generally
in response to pegylated IFN.2The correlations with the
other five ISGs were significant, albeit weak with small
correlation coefficients. This could be explained by the
fact that cleavage of MAVS occurs only in hepatocytes
infected with HCV, whereas activation of ISGs involves
all liver cells because of the paracrine effects of secreted
tion at the single cell will be required to address this issue;
however, this is still a technical challenge. Alternatively,
the weak correlation between cleavage of MAVS and ISG
induction could be explained by MAVS cleavage being
only one of several factors that determine the activation
status of the endogenous IFN system. Other factors with
a possible impact on pre-activation include NS3-4A–me-
diated cleavage of TRIF,16inhibition of IFN regulatory
phatase, a recently identified cellular substrate of the
subsequently showed EVR to therapy with pegylated
IFN-? and ribavirin (Fig. 4H). Given the known corre-
lation between treatment NR and pre-activation of the
of MAVS cleavage in regulating the activation status of
the endogenous IFN system. However, many patients
with cleaved MAVS do not respond to therapy (and vice
Fig. 4. Activation of the endogenous IFN system in the liver and cleavage
of MAVS. (A) Quantification of p-STAT1 nuclear staining in hepatocytes of
correlation to the percentage of cleaved MAVS. Shown are the mean values
with standard error of the mean. One-way analysis of variance analysis
showed a significant difference between the four categories of p-STAT1
staining (P ? 0.023, R2? 0.153) and a significant linear trend (slope ?
?4.58, R2? 0.149, P ? 0.0025). N ? number of patients in each group.
(B-G) Correlation of percentage of cleaved MAVS with the mRNA expression
of six ISGs. There was a significant negative correlation with IFI44L, Viperin,
IFI27, USP18, and STAT1. Sixty-two samples from patients with CHC were
analyzed. (H) CHC patients with an EVR have significantly more cleavage of
MAVS compared with patients with a PNR. Shown are the mean values with
standard error of the mean. The P-values were obtained with the t test. N ?
number of patients in each group.
Fig. 5. Preactivation of the endogenous IFN system and HCV viral
load. Serum (A) or intrahepatic (B) HCV VLs did not significantly differ
between patients showing ?5% (?), 5%-33% (?), 34%-66% (??), or
?66% nuclear (???) p-STAT1 staining in hepatocytes (Kruskal-Wallis
test, P ? 0.26 for serum VL and P ? 0.81 for intrahepatic VL).
1134 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.HEPATOLOGY, April 2010
versa), and quantification of MAVS cleavage in pretreat-
ment biopsy specimens therefore cannot accurately pre-
dict response to treatment. Furthermore, we did not find
a significant correlation of MAVS cleavage with final
treatment outcomes (data not shown).
Not only HCV GT but also serum and intrahepatic
VL significantly correlated with cleavage of MAVS. Pa-
tients with high VL showed more cleavage of MAVS in
the liver (Fig. 2) and might be expected to have a weaker
activation of the endogenous IFN system. Such a correla-
activation of the endogenous IFN system could allow an
increased viral replication, resulting in a high VL. How-
ever, we could not confirm this notion, neither by mea-
suring the activation of the Jak-STAT pathway by
quantification of nuclear p-STAT1 staining (Fig. 5), nor
by correlation analysis between VL and ISG expression
the lack of inverse correlation between baseline VL and
pre-activation. First of all, our analysis with 129 patients
might be underpowered to show a significant correlation
between baseline VL and pre-activation. In our model,
cleavage, and the correlations between VL and MAVS
cleavage as well as between MAVS cleavage and ISG in-
duction are weak. Second, high VL increases the risk of
NR to treatment with pegylated IFN-? and ribavirin by
yet unknown mechanisms.26Because a pre-activation of
the endogenous IFN system also strongly correlates with
NR,2,17,18VL should positively correlate with the activa-
tion status of the endogenous IFN system. However, as
shown in Fig. 5, there is neither a positive nor a negative
correlation between VL and activation of the IFN system
in the liver. Apparently, the effect of high VL on cleavage
of MAVS is abrogated by yet unknown effects of VL on
the induction of the endogenous IFN system, resulting in
the lack of correlation between VL and pre-activation of
the IFN system shown in Figure 5.
The number of infected hepatocytes in CHC has not
been determined unequivocally, because the spread of
HCV infection in the liver is difficult to assess because of
gue that a limited proportion of hepatocytes harbor rep-
licating HCV,33-35others suggest a more widespread
infection at least in some patients.36-38Strikingly, we ob-
served up to 76% MAVS cleavage (Fig. 1B), suggesting
that in some patients HCV infection is widespread, be-
tocytes harboring HCV. This notion is supported by
experiments in which Huh-7 cells harboring HCV repli-
cons were co-cultured with Huh-7 cells expressing the
only in replicon-harboring cells (P.B. and D.M., unpub-
In conclusion, our data demonstrate an important role
of HCV-induced cleavage of MAVS in the interaction
between virus and host. MAVS cleavage can be detected
liver. Patients with high VL and GT 2 and 3 infections
have MAVS cleaved more often and more extensively.
However, the correlation of MAVS cleavage with pre-
activation of the endogenous IFN system and with re-
not strong enough to use this parameter for patient man-
agement. Our results indicate that MAVS is just one of
probably many factors that control virus–host interac-
tions in CHC. Although this is currently debated,39there
tease inhibitors on the innate immune system in the liver,
and they should be studied in the future by analyzing
MAVS cleavage, IFN signaling, and ISG induction in
liver biopsy specimens of patients undergoing such novel
1. Bigger CB, Guerra B, Brasky KM, Hubbard G, Beard MR, Luxon BA, et
al. Intrahepatic gene expression during chronic hepatitis C virus infection
in chimpanzees. J Virol 2004;78:13779-13792.
hepatitis C. Proc Natl Acad Sci U S A 2008;105:7034-7039.
3. Gale M, Jr., Foy EM. Evasion of intracellular host defence by hepatitis C
virus. Nature 2005;436:939-945.
4. Dustin LB, Rice CM. Flying under the radar: the immunobiology of
hepatitis C. Annu Rev Immunol 2007;25:71-99.
5. Chung RT, Gale M Jr, Polyak SJ, Lemon SM, Liang TJ, Hoofnagle JH.
Mechanisms of action of interferon and ribavirin in chronic hepatitis C:
summary of a workshop. HEPATOLOGY 2008;47:306-320.
6. Rehermann B. Hepatitis C virus versus innate and adaptive immune re-
sponses: a tale of coevolution and coexistence. J Clin Invest 2009;119:
7. Seth RB, Sun L, Ea CK, Chen ZJ. Identification and characterization of
MAVS, a mitochondrial antiviral signaling protein that activates NF-kap-
paB and IRF 3. Cell 2005;122:669-682.
8. Meylan E, Curran J, Hofmann K, Moradpour D, Binder M, Barten-
and is targeted by hepatitis C virus. Nature 2005;437:1167-1172.
9. Kawai T, Takahashi K, Sato S, Coban C, Kumar H, Kato H, et al. IPS-1,
an adaptor triggering RIG-I- and Mda5-mediated type I interferon induc-
tion. Nat Immunol 2005;6:981-988.
10. Xu LG, Wang YY, Han KJ, Li LY, Zhai Z, Shu HB. VISA is an adapter
11. Yoneyama M, Kikuchi M, Natsukawa T, Shinobu N, Imaizumi T, Miy-
agishi M, et al. The RNA helicase RIG-I has an essential function in
double-stranded RNA-induced innate antiviral responses. Nat Immunol
12. Fitzgerald KA, McWhirter SM, Faia KL, Rowe DC, Latz E, Golenbock
DT, et al. IKKepsilon and TBK1 are essential components of the IRF3
signaling pathway. Nat Immunol 2003;4:491-496.
HEPATOLOGY, Vol. 51, No. 4, 2010 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.1135
13. Sharma S, tenOever BR, Grandvaux N, Zhou GP, Lin R, Hiscott J. Trig-
gering the interferon antiviral response through an IKK-related pathway.
14. Yoneyama M, Suhara W, Fukuhara Y, Fukuda M, Nishida E, Fujita T.
Direct triggering of the type I interferon system by virus infection: activa-
tion of a transcription factor complex containing IRF-3 and CBP/p300.
EMBO J 1998;17:1087-1095.
15. Lin R, Heylbroeck C, Pitha PM, Hiscott J. Virus-dependent phosphory-
lation of the IRF-3 transcription factor regulates nuclear translocation,
16. Li K, Foy E, Ferreon JC, Nakamura M, Ferreon AC, Ikeda M, et al.
Immune evasion by hepatitis C virus NS3/4A protease-mediated cleavage
17. Chen L, Borozan I, Feld J, Sun J, Tannis LL, Coltescu C, et al. Hepatic
gene expression discriminates responders and nonresponders in treatment
of chronic hepatitis C viral infection. Gastroenterology 2005;128:1437-
18. Asselah T, Bieche I, Narguet S, Sabbagh A, Laurendeau I, Ripault MP, et
al. Liver gene expression signature to predict response to pegylated inter-
C. Gut 2008;57:516-524.
19. Saito T, Owen DM, Jiang F, Marcotrigiano J, Gale M, Jr. Innate immu-
nity induced by composition-dependent RIG-I recognition of hepatitis C
virus RNA. Nature 2008;454:523-527.
and therapeutic control of IFN-beta promoter stimulator 1 during hepa-
titis C virus infection. Proc Natl Acad Sci U S A 2006;103:6001-6006.
21. Blight KJ, McKeating JA, Rice CM. Highly permissive cell lines for sub-
genomic and genomic hepatitis C virus RNA replication. J Virol 2002;76:
22. Moradpour D, Kary P, Rice CM, Blum HE. Continuous human cell lines
inducibly expressing hepatitis C virus structural and nonstructural pro-
teins. HEPATOLOGY 1998;28:192-201.
A. Serum and liver HCV RNA levels in patients with chronic hepatitis C:
correlation with clinical and histological features. Gut 1998;42:856-860.
in the liver of chronic hepatitis C patients by strand-specific semiquanti-
25. Wieland SF, Chisari FV. Stealth and cunning: hepatitis B and hepatitis C
viruses. J Virol 2005;79:9369-9380.
26. Manns MP, McHutchison JG, Gordon SC, Rustgi VK, Shiffman M,
Reindollar R, et al. Peginterferon alfa-2b plus ribavirin compared with
a randomised trial. Lancet 2001;358:958-965.
27. Lehmann M, Meyer MF, Monazahian M, Tillmann HL, Manns MP,
Wedemeyer H. High rate of spontaneous clearance of acute hepatitis C
virus genotype 3 infection. J Med Virol 2004;73:387-391.
28. Harris HE, Eldridge KP, Harbour S, Alexander G, Teo CG, Ramsay ME.
Does the clinical outcome of hepatitis C infection vary with the infecting
hepatitis C virus type? J Viral Hepat 2007;14:213-220.
29. Lagging M, Romero AI, Westin J, Norkrans G, Dhillon AP, Pawlotsky
JM, et al. IP-10 predicts viral response and therapeutic outcome in diffi-
of interferon regulatory factor-3 by the hepatitis C virus serine protease.
31. Otsuka M, Kato N, Moriyama M, Taniguchi H, Wang Y, Dharel N, et al.
Interaction between the HCV NS3 protein and the host TBK1 protein
leads to inhibition of cellular antiviral responses. HEPATOLOGY 2005;41:
tyrosine phosphatase. HEPATOLOGY 2009;49:1810-1820.
hepatitis C virus: a critical appraisal. J Hepatol 1996;24:43-51.
34. Vona G, Tuveri R, Delpuech O, Vallet A, Canioni D, Ballardini G, et al.
Intrahepatic hepatitis C virus RNA quantification in microdissected hepa-
tocytes. J Hepatol 2004;40:682-688.
ton M, et al. Hepatitis C virus antigen in hepatocytes: immunomorpho-
logic detection and identification. Gastroenterology 1992;103:622-629.
36. Gosalvez J, Rodriguez-Inigo E, Ramiro-Diaz JL, Bartolome J, Tomas JF,
Oliva H, et al. Relative quantification and mapping of hepatitis C virus by
in situ hybridization and digital image analysis. HEPATOLOGY 1998;27:
37. Agnello V, Abel G, Knight GB, Muchmore E. Detection of widespread
hepatocyte infection in chronic hepatitis C. HEPATOLOGY 1998;28:573-
hepatic tissue and its correlation with liver disease. J Virol 2000;74:944-
39. Liang Y, Ishida H, Lenz O, Lin TI, Nyanguile O, Simmen K, et al. Anti-
viral suppression vs restoration of RIG-I signaling by hepatitis C protease
and polymerase inhibitors. Gastroenterology 2008;135:1710-1718.
1136 BELLECAVE, SARASIN-FILIPOWICZ, ET AL.HEPATOLOGY, April 2010