Activation of Toll-like receptor 3 impairs the dengue virus serotype 2 replication through induction of IFN-β in cultured hepatoma cells.

Abstract

Toll-like receptors (TLRs) play an important role in innate immunity against invading pathogens. Although TLR signaling has been indicated to protect cells from infection of several viruses, the role of TLRs in Dengue virus (DENV) replication is still unclear. In the present study, we examined the replication of DENV serotype 2 (DENV2) by challenging hepatoma cells HepG2 with different TLR ligands. Activation of TLR3 showed an antiviral effect, while pretreatment of other TLR ligands (including TLR1/2, TLR2/6, TLR4, TLR5 or TLR7/8) did not show a significant effect. TLR3 ligand poly(I:C) treatment prior to viral infection or simultaneously, but not post-treatment, significantly down-regulated virus replication. Pretreatment with poly(I:C) reduced viral mRNA expression and viral staining positive cells, accompanying an induction of the type I interferon (IFN-β) and type III IFN (IL-28A/B). Intriguingly, neutralization of IFN-β alone successfully restored the poly(I:C)-inhibited replication of DENV2. The poly(I:C)-mediated effects, including IFN induction and DENV2 suppression, were significantly reversed by IKK inhibitor, further suggesting that IFN-β is the dominant factor involved in the poly(I:C) mediated antiviral effect. Our study presented the first evidence to show that activation of TLR3 is effective in blocking DENV2 replication via IFN-β, providing an experimental clue that poly(I:C) may be a promising immunomodulatory agent against DENV infection and might be applicable for clinical prevention.

Full text

Page 1

Activation of Toll-Like Receptor 3 Impairs the Dengue
Virus Serotype 2 Replication through Induction of IFN-b
in Cultured Hepatoma Cells
Zhaoduan Liang1,2., Siyu Wu1,2., Yuye Li1,2, Li He1,2, Minhao Wu1,2, Lifang Jiang2,3, Lianqiang Feng1,2,
Ping Zhang1,2*, Xi Huang1,2*
1Department of Immunology, Institute of Immunology, Institute of Human Virology, Zhongshan School of Medicine, Sun Yat-sen University, Guangzhou, China, 2Key
Laboratory of Tropical Diseases Control (Sun Yat-sen University), Ministry of Education, Guangzhou, China, 3Department of Microbiology, Zhongshan School of Medicine,
Sun Yat-sen University, Guangzhou, China
Abstract
Toll-like receptors (TLRs) play an important role in innate immunity against invading pathogens. Although TLR signaling has
been indicated to protect cells from infection of several viruses, the role of TLRs in Dengue virus (DENV) replication is still
unclear. In the present study, we examined the replication of DENV serotype 2 (DENV2) by challenging hepatoma cells
HepG2 with different TLR ligands. Activation of TLR3 showed an antiviral effect, while pretreatment of other TLR ligands
(including TLR1/2, TLR2/6, TLR4, TLR5 or TLR7/8) did not show a significant effect. TLR3 ligand poly(I:C) treatment prior to
viral infection or simultaneously, but not post-treatment, significantly down-regulated virus replication. Pretreatment with
poly(I:C) reduced viral mRNA expression and viral staining positive cells, accompanying an induction of the type I interferon
(IFN-b) and type III IFN (IL-28A/B). Intriguingly, neutralization of IFN-b alone successfully restored the poly(I:C)-inhibited
replication of DENV2. The poly(I:C)-mediated effects, including IFN induction and DENV2 suppression, were significantly
reversed by IKK inhibitor, further suggesting that IFN-b is the dominant factor involved in the poly(I:C) mediated antiviral
effect. Our study presented the first evidence to show that activation of TLR3 is effective in blocking DENV2 replication via
IFN-b, providing an experimental clue that poly(I:C) may be a promising immunomodulatory agent against DENV infection
and might be applicable for clinical prevention.
Citation: Liang Z, Wu S, Li Y, He L, Wu M, et al. (2011) Activation of Toll-Like Receptor 3 Impairs the Dengue Virus Serotype 2 Replication through Induction of
IFN-b in Cultured Hepatoma Cells. PLoS ONE 6(8): e23346. doi:10.1371/journal.pone.0023346
Editor: Dong-Yan Jin, University of Hong Kong, Hong Kong
Received April 21, 2011; Accepted July 15, 2011; Published August 4, 2011
Copyright: ? 2011 Liang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported in part by NSFC (National Natural Science Foundation of China) U0832006, 30972763, GRPCRG (Guangdong Recruitment
Program of Creative Research Groups), Doctoral Fund of Ministry of Education of China 20100171110047, Guangdong Natural Science Foundation
10251008901000013. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: huangxi6@mail.sysu.edu.cn (XH); zhangp36@mail.sysu.edu.cn (PZ)
. These authors contributed equally to this work.
Introduction
Dengue virus (DENV), a group of prevalent arthropod-borne
viruses, belongs to the Flavirvirus genus of the family Flaviviridae.
There are four related but distinct serotype viruses (DENV-1 to 4)
[1]. While most DENV infections are asymptomatic or cause a
self-limited Dengue fever (DF), some infections lead to severe and
potentially lethal diseases, such as Dengue hemorrhagic fever
(DHF) and Dengue shock syndrome (DSS) [1,2]. Dengue becomes
a leading infectious disease with more than 50 million cases of DF,
500 thousand cases of DHF and 20 thousand deaths each year
world wide [1].
So far, DENV pathogenesis and host immune response to
virus infection are largely unclear, and no effective vaccines or
antiviral drugs are available for Dengue diseases. Given innate
immunity especially interferons (IFNs) plays an important role in
limiting viral replication during DENV primary infection,
amplifying the innate immune response might be an effective
strategy to defend DENV infection. There are three type IFNs,
the type I IFNs, including 13 IFN-as and a single IFN-b, IFN-v,
IFN-e or IFN-k, are well-known as antiviral IFNs [3]. The type
I IFNs are broadly expressed in different cell types, including
lymphocytes, macrophages, fibroblasts, endothelial cells, hepa-
toma cells and so on. The type II IFN-c, also known as immune
IFN, is expressed in natural killer (NK) or activated T cells. The
type III IFN family comprises three subtypes, i.e. IL-28A/B and
IL-29, which are co-produced with IFN-b and share many
functional characteristics with type I IFNs [4]. After synthesis
and secretion, IFNs bind to cell surface receptors and activate
downstream signaling pathways via Janus kinase (JAK), tyrosine
kinase (TYK) 2, signal transducers and activators of transcrip-
tion (STAT) 1/2 and interferon regulatory factor (IRF) 9,
resulting in an induction of more than 300 interferon stimulated
genes (ISGs) [3].
The type I and II IFNs have been showed to restrict the
propagation of DENV in both in vitro and in vivo models. In in vitro
cultured cells, pretreatment with either type I IFNs (IFN-a,-b) or
type II IFN (IFN-c) were effective in inhibiting DENV infection
[5]. And in in vivo murine model, the deficiency of IFN-a, -b, -c
receptors led to a significantly higher lethality upon intraperitoneal
PLoS ONE | www.plosone.org1August 2011 | Volume 6 | Issue 8 | e23346

Page 2

inoculation of DENV2 [6]. Intriguingly, a recent report indicated
that the absence of type I IFN expression in DENV2 infected
human monocyte-derived dendritic cells was the major reason that
DC cells failed to prime T cells [7]. In terms of type III IFN, its
expression pattern and role in DENV infection have not been
investigated yet.
IFNs and other proinflammatory cytokines are rapidly induced
by innate immune system in responding to pathogen invasion. The
innate signal is originated from recognition of pathogen-associated
molecular patterns (PAMPs) on pathogens by pattern recognition
receptors (PRRs). For example, the PRRs to recognize DENV
PAMPs include both endosomally-localized Toll-like receptors
(TLR) 3, 7, 8 and cytoplamsic RIG-I/MDA5 [8,9,10]. Then, the
activated signal is transmitted through adaptor proteins including
myeloid differentiation primary response gene 88 (MyD88), Toll/
IL-1 receptor domain-containing adapter-inducing interferon-b
(TRIF) or interferon-beta promoter stimulator (IPS)-1. Subse-
quently, downstream phosphorylation cascade of the IkB kinase
family (IKK-a,-b,-e) is activated, resulting in an induction of IFNs
and cytokines [3].
An important family of PRRs, TLRs, includes 10 members in
humans. TLRs are responsible for detecting PAMPs of various
pathogens, such as bacterial LPS (TLR4), flagellin (TLR5),
unmethylated DNA (TLR9); viral dsRNA(TLR3) or ssRNA
(TLR7/8), etc [11,12,13]. TLRs are expressed in both immune
cells and non-immune cells, such as epithelial cells and
hepatoma cells [12,13]. Most TLRs signal through adaptor
MyD88, exceptthat TLR3
Activation of the innate immune signaling by TLR ligands has
been broadly utilized for the prevention and therapy of
infectious diseases [11,13]. So far, the protective effects of
TLRs in cultured cells or animal models against a variety of
viruses have been well documented. In both in vivo murine
models and in vitro cultured cells, HBV replication was inhibited
by TLR2, 3, 4, 5, 7, 9 agonists in a type I IFN-dependent
manner [14,15,16]. Treatment with TLR3, 5, 7, 9 ligands
significantly inhibited Herpes virus type 2 (HSV-2) replication in
mice model or primary genital epithelial cells [17,18]. And
stimulation with TLR3 ligands inhibited HIV amplification in
dendritic cells and macrophages [19,20], HCV replication in
human heptocytes [21], replication of influenza virus [22] and
respiratory syncytial virus (RSV) [23], etc. Regarding to the
effect of TLR activation on the DENV replication, only a few
studies on the TLR4 activation have been reported while their
conclusions were illusive. One group found that LPS inhibited
DENV infection in monocytes/macrophage by blocking virus
entry [24], while another group found that LPS pretreatment
did not significantly reduce DENV replication in HepG2 cells
[25]. Interestingly, DENV was found to counter the TLR
activating pathway, further implying that TLR signaling might
have some roles in blocking DENV infection [26].
The present study aimed to investigate the role of innate
immune response in non-immune cells but actively infected by
DENV. Our approach was to screen the effect of individual TLR
ligands on DENV serotype 2 (New Guinea C strain, NGC)
replication in hepatoma cells HepG2, a commonly-used DENV
permissive cell line [27] with most TLR genes expressed [14,28].
And we found that these TLR ligands play differential roles in the
DENV2 infection. IFN-b induced by poly(I:C) was identified to be
the dominant antiviral agent. For the first time, we showed that the
protective role of TLR3 activation on DENV replication was
dependent on the time of treatment. These results provide
substantial evidence that poly(I:C) might be a promising
prevention agent against Dengue diseases.
signals through TRIF[12].
Results
TLR ligands played differential roles in DENV2 replication
Previous study have shown that TLRs, including TLR2, TLR3,
TLR6, TLR9, are consistently expressed in HepG2 cells, while
TLR1, TLR4, TLR5, TLR8 are expressed at a very weak level
[28]. To validate the expression of TLRs in HepG2 cells,
conventional RT-PCR was carried out. The result showed that
expression of most TLRs, including TLR1 through TLR8 (except
for TLR2), were detected (Fig. 1A). To further determine whether
TLR signaling pathway is activated, the induction of proinflam-
matory cytokine IL-6 and TNF-a in each TLR ligands pretreated
HepG2 cells were measured by real-time PCR. Surprisingly, only
treatment of TLR3 ligand significant induced both IL-6 and TNF-
a in HepG2 cells, but not other TLR ligands, including PAM3
(TLR1/2), PAM2 (TLR2/6), LPS (TLR4), FLA-ST(TLR5) or
R848 (TLR7/8) (Fig. 1B).
To further test whether treatment of TLR ligands has an effect
in the DENV2 replication, HepG2 cells were treated with
individual TLR ligands followed by DENV2 infection. And the
extracellular virus yields in infected cells were examined by
TCID50 assay. As shown in Fig. 1A, about 104PFU/ml viral
particles were produced in mock treated cells, confirming that
HepG2 cells are permissive to DENV. Among the TLR ligands,
TLR3 ligand poly(I:C) pretreatment significantly decreased the
DENV2 yields (by 90%), while all the other TLR ligands,
including ligands of TLR1/6, TLR2/6, TLR4, TLR5 or TLR7/8
had no or very weak effect on the DENV2 production.
The extracellular DENV2 yields were also tested among three
groups of HepG2 cells that were challenged with poly(I:C) at
different time points, namely 9 h prior to, simultaneously, or 9 h
post virus inoculation. Our data showed that simultaneous
treatment of poly(I:C) with DENV2 infection significantly
decreased the production of virus particles by 90% (Fig. 1B). By
contrast, post-treatment of poly(I:C) totally failed to execute a
protective effect, resulting in a similar virus amount as non-treated
control.
Pretreatment of poly(I:C) suppressed the DENV2
replication
Because pre- and simultaneous treatment with poly(I:C) resulted
inasimilaranti-DENV2effect,thepoly(I:C)pretreatmentwashence
chosen for all following investigations. To determine whether
poly(I:C) inhibits DENV2 production by blocking DENV2 replica-
tion at transcriptional or posttranscriptional level, the expression of
viral mRNA or protein in poly(I:C)-treated HepG2 cells have been
tested. The kinetic viral mRNA levels of DENV2 were progressively
increased in both control (white bars) and poly(I:C) pretreated
HepG2 cells (black bars) after virus infection at 6, 12 and 24 h p.i
(Fig. 2A). However, the increasing extent of viral mRNA copies was
drastically attenuated by 78% (6 h p.i.), 85% (12 h p.i.), and 85%
(24 h p.i.) respectively, in poly(I:C)- vs mock-pretreated cells (6 h:
P,0.001; 12 h: P,0.05; 24 h: P,0.05). Moreover, the prM protein
of DENV2 was examined by immunofluorescence microscopy using
the anti-prM antibody. Approximate 25% HepG2 cells were prM
positive at 24 h after infection with DENV2 virus, while there were
only 4% cells positively stained in poly(I:C) pretreated cells
(p,0.001) (Fig. 2B, 2C). There was no positive staining in the
mock-infected control cells.
Pretreatment of poly(I:C) induced type I and III IFN
expression
To investigate the mechanism by which poly(I:C) pretreatment
inhibits the DENV2 replication, expression of IFNs, including type
TLR3 Activation Impairs Dengue Virus Replication
PLoS ONE | www.plosone.org2August 2011 | Volume 6 | Issue 8 | e23346

Page 3

I IFN (represented by IFN-b), type II (IFN-c) and type III (IL-
28A/B) were examined by real-time PCR. IFN-c expression was
not altered by poly(I:C) treatment (data not shown). However, the
expression levels of IFN-b and IL-28A/B were significantly
increased by 30- or 400- fold respectively (P,0.001) (0 h p.i.,
Fig. 3A, 3B). Moreover, the IFN-b and IL-28A/B mRNA
expression levels gradually decreased after removal of poly(I:C)
(6, 12, and 24 h p.i., Fig. 3A, 3B, black bars). The infection of
DENV2 did not elicit an additional induction of IFN-b and IL-
28A/B within 24 h p.i., in either mock- or poly(I:C)- treated
HepG2 cells (Fig. 3A, 3B, white bars and hatched bars).
Poly(I:C)-inhibited virus replication was mediated
through IFN-b
Although the expression of IL-28A/B was induced by poly(I:C)
treatment, pretreatment of IL-28A/B alone was not able to
effectively protect HepG2 cells from DENV2 infection (data not
shown). To test the role of IFN-b, the viral replication levels in the
presence of a neutralizing antibody against IFN-b were examined.
Surprisingly, the neutralization of IFN-b only was sufficient to
restore the viral replication. In Fig. 4A and 4B, the viral mRNA
levels in poly(I:C) pretreated cells and the extracelluar virus yields
(hatched bars) were completely restored by the IFN-b neutralizing
antibody, reaching a similar level to DENV2 infection alone (white
bars). The viral replication levels recovered by neutralizing
antibody were about 17-fold higher than poly(I:C) pretreated cells
(black bars).
Poly(I:C)-mediated IFN production and virus replication
suppression were dependent on IKK
To further establish the co-relationship between the IFN-b and
DENV2 replication mediated by poly(I:C), we compared the IFN-
b expression and viral replication levels in the cells treated with an
IKK inhibitor (BMS-345541) before poly(I:C) treatment and
DENV2 infection. In the presence of IKK inhibitor (hatched bars,
Fig. 5A), induction of IFN-b by poly(I:C) was significantly
decreased by 78.5% compared to poly(I:C) pretreated cells
(P,0.001)(black bars, Fig. 5A) as predicted. On the other hand,
percentage of positive stained cells (approximate 15%) was
significantly recovered by addition of IKK inhibitor (panel d,
Fig. 5B; hatched bar, Fig. 5C) compared to the nontreated control
(panel c, Fig. 5B; black bar, Fig. 5C). Moreover, the attenuated
Figure 1. Effect of TLR ligands on DENV2 replication in HepG2 cells. (A) Endogenous TLR expression in HepG2 cells. (B) Induction of IL-6 and
TNF-a in response to TLR ligands. HepG2 cells were pretreated with individual TLR ligands for 9 h. (C) Extracellular virus yields. HepG2 cells were
pretreated with individual TLR ligands and infected by DENV2 for 24 h. (D) Extracellular virus yields. HepG2 cells were pretreated with poly(I:C) at 9 h
prior to, simultaneously (simu), or 9 h post DENV2 infection. Supernatants were harvested at 24 h p.i. and titered on C6/36 cells. Error bars represent
the standard error of mean from the average of three experiments. Student’s t test, **, p,0.01; ***, p,0.001.
doi:10.1371/journal.pone.0023346.g001
TLR3 Activation Impairs Dengue Virus Replication
PLoS ONE | www.plosone.org3August 2011 | Volume 6 | Issue 8 | e23346

Page 4

extracellular virions (black bar, Fig. 5B) were also significantly
recovered by IKK inhibitor (hatched bar, Fig. 5D), and it was only
50% less than non-treated control (white bar, Fig. 5D).
Discussion
TLRs are an important family of PRRs to detect PAMPs of
invading pathogen and initialize innate immune response. TLRs
and their signaling pathways are emerging as novel targets of
therapeutics [29]. In the present study, we screened the effects of
TLRs in the DENV2 replication in a widely used DENV
permissive cell line, hepatoma HepG2. We demonstrated that
TLR3 ligand poly(I:C) has a protective effect in DENV2
replication, in which IFN-b plays a determinant role.
Differential effects of TLR ligands in suppressing DENV2
replication were observed. Our PCR result found that most of
TLRs, except for TLR2, are endogenously expressed in the
HepG2 cells, which was slightly different from previous reports
showing other TLRs, such as TLR1 and TLR8 are not or weakly
expressed while TLR2 is expressed [14,28]. The difference of
TLR expression might be due to the different resources of HepG2
cells or the efficacy of PCR primers. Strikingly, despite of the
presence of these TLRs, the IL-6 and TNF-a production were not
induced by most TLR ligands (except for TLR3) treatment,
indicating that these TLR signaling pathways were not activated
by their ligands in HepG2 cells. The absence of TLR activation is
probably a major reason that these TLR ligands fail to protect cells
from DENV2 infection. Our data found that LPS (TLR4 ligand)
pretreatment did not reduce the DENV infection in HepG2 cells,
which agrees with previous work by Cabrera-Hernandez et al.
[25], but differs from Chen’s finding that LPS inhibited DENV
infection by 1–2 log in TLR-abundant immune cells (primary
monocytes/macrophage) [24]. These contradictory observations
could be attributed to the different TLR4 abundances and
Figure 2. Pretreatment of poly(I:C) suppresses the DENV2 replication in HepG2 cells. HepG2 cells were mock- or pretreated with 5 mg poly
(I:C) for 9 h, then infected with DENV2. (A) Real-time PCR to measure viral mRNA levels. Total RNAs were harvested at 2, 6, 12, and 24 h p.i., and used
for real-time RT-PCR. (B) Immunofluorescence microscopy. Infected cells were fixed at 24 h p.i. and incubated with DENV2 prM antibody. (C)
Percentages of positive-stained cells determined by PicCnt 1006. Error bars represent the standard error of mean from the average of three
experiments. Student’s t test, *, p,0.05; ***, p,0.001.
doi:10.1371/journal.pone.0023346.g002
TLR3 Activation Impairs Dengue Virus Replication
PLoS ONE | www.plosone.org4August 2011 | Volume 6 | Issue 8 | e23346

Page 5

activation extents in different experimental cells. Our in vitro result
was further validated by a latest study showing that in vivo
cocurrent activation of TLR3/7/8 decreased viremia and
melioration of DENV1 infection via increased inflammatory and
humoral responses [30].
Protective role of type I and II IFNs has been well established in
inhibiting DENVs infection [5]. Our study showed that poly(I:C)
pretreatment led to an induction of type I and III IFNs in HepG2
cells, but not type II IFN, which is consistent with the fact that
IFN-c is majorly expressed in NK or activated T cells [3]. To
investigate whether anti-DENV effect by TLR3 ligand is IFN-
dependent, we measured the viral replication levels in the presence
of IFN-b neutralizing antibody or IKK inhibitor. Our results
demonstrated that IFN-b neutralizing antibody was sufficient to
recover the viral replication suppressed by poly(I:C), which
indicates that IFN-b is the determinant factor in the TLR3
protective effect. To be noted, the poly(I:C)-induced IFN-b
expression was not totally abrogated by IKK inhibitor and the
DENV2 infection was not fully recovered, most probably because
IKK inhibitor used here is designed to effectively inhibit IKK-1,
but not other IKKs, such as IKKe, a key regulatory protein in IFN
signaling pathway via phosphorylating IRF-3 [3]. Apparently, the
poly(I:C) inhibitory mechanism in our model is distinct from one
of LPS in monocytes or macrophages model, which directly
interfered with the virus entry in a CD14-dependent manner [24].
Furthermore, our real-time PCR data showed that the viral
mRNA levels were decreased by poly(I:C) pretreatment, indicating
that poly(I:C) protective effect might be executed as early as viral
transcription stage. It is well-known that antiviral effects of IFN-b
are mediated by ISG proteins, which cooperate to combat virus
replication at each step [3,31]. Among these ISGs, 29,59-
oligoadenylate synthetase (OAS) 1, together with its downstream
effector RNase L, play an antiviral role at transcription level.
OAS1 and RNase L inhibit a broad range of RNA viruses by
degradation of viral RNA [32]. In a recent study, the infection of
DENV2 PL046 strain was reduced by overexpression of OAS or
RNase L, and enhanced by knockdown of RNase L in human cell
lines [33]. In our assay, the OAS expression was upregulated by
10-fold with poly(I:C) pretreatment (data not shown), suggesting
that OAS and RNase L are, at least partially, responsible for the
TLR3-mediated protective effect, while engagement of other
ISGs, such as protein kinase regulated by dsRNA (PKR) and Mx1,
could not be excluded [3,31].
Our study was the first to detect an induction of type III IFNs in
HepG2 cells in response to poly(I:C) challenge. In vitro, type III
IFNs were shown to have a broad antiviral activity against HBV,
Figure 3. Expression of IFNs in mock-treated or poly(I:C)
pretreated HepG2 cells. After pretreatment with poly(I:C) for 9 h,
HepG2 cells were infected with DENV2 at an MOI 1. The cells were
harvested at 0, 6, 12, and 24 h p.i. for RNA extraction and real-time RT-
PCR for IFN-b (A) or IL-28A/B (B). Error bars represent the standard error
of mean from the average of three experiments. Student’s t test, **,
p,0.01; ***, p,0.001.
doi:10.1371/journal.pone.0023346.g003
Figure 4. DENV2 replication levels in IFN-b neutralizing
antibody treated HepG2 cells. 500 unit/ml IFN-b neutralizing
antibody was added at 1 h before poly(I:C) treatment and kept through
poly(I:C) treatment and virus infection. (A) Real-time PCR detecting the
expression of viral mRNA; (B) Extracellular viral production determined
by TCID50. Supernatants were harvested at 24 h p.i. and titered on C6/
36 cells. Error bars represent the standard error of mean from the
average of three experiments. Student’s t test, **, p,0.01; ***, p,0.001.
doi:10.1371/journal.pone.0023346.g004
TLR3 Activation Impairs Dengue Virus Replication
PLoS ONE | www.plosone.org5August 2011 | Volume 6 | Issue 8 | e23346