of November 13, 2017.This information is current as

Antiviral Immunity against Dengue VirusThe Roles of IRF-3 and IRF-7 in Innate

Andrew D. Luster and Sujan ShrestaHui-Wen Chen, Kevin King, Jui Tu, Marisa Sanchez,

ol.1300799http://www.jimmunol.org/content/early/2013/09/14/jimmun

published online 16 September 2013J Immunol 

MaterialSupplementary

9.DC1http://www.jimmunol.org/content/suppl/2013/09/16/jimmunol.130079

        average*  

4 weeks from acceptance to publicationSpeedy Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2013 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on N

ovember 13, 2017

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

The Roles of IRF-3 and IRF-7 in Innate Antiviral Immunityagainst Dengue Virus

Hui-Wen Chen,*,1 Kevin King,* Jui Tu,* Marisa Sanchez,* Andrew D. Luster,† and

Sujan Shresta*

We investigated the roles of IFN regulatory factor (IRF)-3 and IRF-7 in innate antiviral immunity against dengue virus (DENV).

Double-deficient Irf-32/272/2 mice infected with the DENV2 strain S221 possessed 1,000–150,000 fold higher levels of viral RNA

than wild-type and single-deficient mice 24 h postinfection (hpi); however, they remained resistant to lethal infection. IFN-a/b was

induced similarly in wild-type and Irf-32/2 mice post–DENV infection, whereas in the Irf-72/2 and Irf-32/272/2 mice, significantly

low levels of IFN-a/b expression was observed within 24 hpi. IFN-stimulated gene induction was also delayed in Irf-32/272/2 mice

relative to wild-type and single-deficient mice. In particular, Cxcl10 and Ifna2 were rapidly induced independently of both IRF-3

and IRF-7 in the Irf-32/272/2 mice with DENV infection. Higher levels of serum IFN-g, IL-6, CXCL10, IL-8, IL-12 p70, and TNF

were also observed in Irf-32/272/2 mice 24 hpi, at which time point viral titers peaked and started to be cleared. Ab-mediated

blockade experiments revealed that IFN-g, CXCL10, and CXCR3 function to restrict DENV replication in Irf-32/272/2 mice.

Additionally, the IFN-stimulated genes Cxcl10, Ifit1, Ifit3, andMx2 can be induced via an IRF-3– and IRF-7–independent pathway

that does not involve IFN-g signaling for protection against DENV. Collectively, these results demonstrate that IRF-3 and IRF-7

are redundant, albeit IRF-7 plays a more important role than IRF-3 in inducing the initial IFN-a/b response; only the combined

actions of IRF-3 and IRF-7 are necessary for efficient control of early DENV infection; and the late, IRF-3– and IRF-7–indepen-

dent pathway contributes to anti-DENV immunity. The Journal of Immunology, 2013, 191: 000–000.

Dengue virus (DENV) is a mosquito-borne pathogen thatposes a serious threat in the tropical and subtropicalregions of the world. Infection with one of the four

serotypes of DENV (DENV1–4) causes extensive morbidity andmortality. Approximately 400 million people are infected each year,and 2.5 billion people are at risk for infection in endemic areas,mainly in Southeast Asia, the Pacific, and the Americas (1). Theclinical manifestations of DENV infection range from mild febrileillness to severe symptoms including dengue hemorrhagic feverand dengue shock syndrome (2). Currently, there is no licensedvaccine or antiviral treatment available for DENV (3).Despite the significant prevalence of DENVworldwide, the virus–

host interactions that determine the viral pathogenesis remain

unclear. The short-course and self-limiting febrile symptom ob-served in most DENV-contracted cases suggest a key role of in-nate immune defenses in controlling DENV infection at the earlystage. Among the intrinsic antiviral factors of the host, IFNs areinvolved in numerous initial responses against viral infections.Accordingly, studies with experimental DENV infection in micehave demonstrated a critical role for both type I and II IFNs in thehost defense against DENV (4, 5). In particular, double-deficient129/Sv mice lacking type I and II IFN receptors (AG129 mice) (6,7), STAT1 and STAT2 (STAT12/2/22/2 mice) (8, 9), or STAT1and type I IFN receptor (STAT12/2/IFNAR2/2) are highly sen-sitive to DENV infection and disease (9). Single-deficient 129/Svor C57BL/6 mice lacking type I IFN receptor (A129 or AB6 mice)are also sensitive to DENV infection, albeit they develop diseaseupon higher viral challenge doses than those required for thedouble-deficient animals lacking components of both type I and IIIFN receptor signaling (6, 10). In contrast, single-deficient micelacking type II IFN receptor have a nearly normal resistance toDENV infection and disease (4, 5), highlighting the critical role oftype I IFN receptor signaling in host defense against DENV.Type I IFN responses against DENV are triggered through the

viral RNA binding of pattern-recognition receptors (RIG-I/MDA5)and downstream MAVS/IPS-1/VISA/Cardif-dependent signaling incultured fibroblasts (11). Consistent with this study, we observeda delayed type I IFN production in MAVS2/2 mice infected withthe DENV2 strain S221, indicating that MAVS regulates the initialtype I IFN response during DENV infection (10). Previously in thismurine model of experimental DENV infection, we have shown thatthe early, highly efficient type I IFN response requires the combinedaction of STAT1 and STAT2, and each STAT pathway can functionindependently to induce type I IFNs and limit viral replication laterin infection (9). At present, the signaling mechanisms that are down-stream of MAVS but upstream of STAT1 or STAT2 for the inductionof type I IFN responses against DENV are as yet to be defined.

*Division of Vaccine Discovery, La Jolla Institute for Allergy and Immunology, LaJolla, CA 92037; and †Division of Rheumatology, Allergy and Immunology, Centerfor Immunology and Inflammatory Diseases, Massachusetts General Hospital, HarvardMedical School, Boston, MA 02115

1Current address: School of Veterinary Medicine, National Taiwan University, Taipei,Taiwan.

Received for publication March 22, 2013. Accepted for publication August 3, 2013.

This work was supported by National Institutes of Health Grants U01 AI082185, U54AI057157 from the Southeast Regional Center of Excellence for Emerging Infectionsand Biodefense and National Institute of Allergy and Infectious Diseases ContractHHSN272200900042C.

The PCR array data presented in this article have been submitted to the NationalCenter for Biotechnology Information’s Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE49871) under accession number GSE49871.

Address correspondence and reprint requests to Dr. Sujan Shresta, Division of Vac-cine Discovery, La Jolla Institute for Allergy and Immunology, 9420 Athena Circle,La Jolla, CA 92037. E-mail address: sujan@liai.org

The online version of this article contains supplemental material.

Abbreviations used in this article: DENV, dengue virus; GE, genomic equivalent; hpi,hours postinfection; IFNAR, type I IFN receptor; IRF, IFN regulatory factor; ISG,IFN-stimulated gene; qRT-PCR, quantitative RT-PCR; WNV, West Nile virus.

Copyright� 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1300799

Published September 16, 2013, doi:10.4049/jimmunol.1300799 by guest on N

ovember 13, 2017

http://ww

w.jim

munol.org/

Dow

nloaded from

IFN regulatory factors (IRF) 3 and 7 are primary transcriptionalfactors downstream of MAVS signaling and regulate the type I IFNresponse after RNA virus infections (12, 13). In this study, wetherefore investigated the roles of IRF-3 and IRF-7 in innate hostimmunity against DENV. Mice deficient in IRF-3, IRF-7, or bothIRF-3 and IRF-7 were infected with the DENV2 strain S221 andexamined at the virologic and immunologic level. We observe thatefficient early control of viral replication requires the combinedaction of IRF-3 and IRF-7. However, each pathway can functionindependently to limit the initial viral replication. Moreover, eventhe combined absence of IRF-3 and IRF-7 is not sufficient to in-duce disease, revealing a role for the IRF-3– and IRF-7–inde-pendent pathway in innate antiviral immunity against DENV.

Materials and MethodsCells and viruses

C6/36 cells (Aedes albopictus mosquito cells) were maintained in Leibo-vitz’s L15 medium (Invitrogen, Carlsbad, CA) supplemented with 10%FBS (Gemini Bio Products, Woodland, CA), penicillin, streptomycin, andHEPES (all from Invitrogen) at 28˚C in the absence of CO2. S221 is a pla-que-purified virus strain derived from the Taiwanese DENV2 clinical isolatePL046 (obtained from Dr. Huan-Yao Lei, National Cheng Kung University,Taiwan). The virus was propagated in C6/36 cells and quantified based ongenomic equivalent (GE) by real-time RT-PCR as previously described(14). There are ∼5 3 104 GE per PFU for a fixed batch of S221 usedthroughout this study, based on a plaque assay on baby hamster kidney cells.

Mice and infections

C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor,ME). Irf-32/2, Irf-72/2, and Irf-32/272/2 breeder mice were provided byDr. Tadatsugu Taniguchi (The University of Tokyo) via Dr. Michael Dia-mond (Washington University). IFNAR2/2 mice were obtained fromDr. Wayne Yokoyama (Washington University, St. Louis, MO) via Dr. CarlWare (La Jolla Institute for Allergy and Immunology). All knockout miceare on the C57BL/6 genetic background. Mice were infected i.v. with 1011

GE DENV S221. All mice experiments were approved by the Animal CareCommittee at La Jolla Institute for Allergy and Immunology.

Measurement of DENV burden in mice

Mice were euthanized via isoflurane inhalation at different time pointspostinfection and subsequently perfused with 30 ml PBS. Blood wascollected via cardiac puncture, and viral RNA in serum was extracted usingthe QIAamp viral RNA Mini kit (Qiagen). Tissues were harvested intoRNAlater (Qiagen) and homogenized in RLT lysis buffer (Qiagen) for 3 minusing Tissuelyser (Qiagen) according to the manufacturer’s instruction.Total RNA in tissues was extracted using the RNeasy Mini kit (Qiagen).Extracted RNA was stored at 280˚C until analysis. Quantitative RT-PCR(qRT-PCR) of DENV viral RNA in serum and tissues was performed aspreviously described (14). A standard curve was generated with knownconcentrations of DENV2 or 18S RNA. Virus titers in serum were ex-pressed as the log10 DENV2 GE per milliliter, and those in tissues wereexpressed as log10 DENV2 GE relative to 18S. Plaque assays of virus inserum samples were performed as previously described (7).

qRT-PCR of type I IFN mRNA

Quantitation of IFN-a and -b RNA was performed by qRT-PCR usingpreviously described primers and probes (15).

Immunohistochemistry

Mice were sacrificed with isoflurane inhalation and perfused with 30 mlPBS. Spleens were embedded in OCT compound (Sakura, Torrance, CA)and stored at280˚C. Cryosections (6 mm) were cut and fixed for 10 min inacetone, followed by 8 min in 1% paraformaldehyde (EMS, Hatfield, PA)in 100 mM dibasic sodium phosphate containing 60 mM lysine and 7 mMsodium periodate (pH 7.4) at 4˚C. Sections were blocked by 5% normalgoat serum (Invitrogen) in PBS for 10 min and stained with affinity-purified rabbit polyclonal anti-DENV NS3 Ab (obtained from NovartisInstitute for Tropical Diseases, Singapore) for 4 h at room temperature.After washes, sections were stained with Dylight 649 labeled goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) for1 h. Nuclei were counterstained using DAPI (Invitrogen).

Measurement of cytokines and chemokine by ELISA

Serum levels of type I IFN were analyzed using the VeriKine Mouse IFN aELISA Kit and Mouse IFN b ELISA Kit (PBL InterferonSource, Piscat-away, NJ). The IFN-a ELISA kit detects all subtypes of IFN-a accordingto the manufacturer. Serum levels of type III IFN (IFN l) were examinedby Mouse IL-28 Platinum ELISA (eBioscience, San Diego, CA). Serumlevels of CXCL10 were measured by Mouse IP-10 Platinum ELISA(eBioscience), and other cytokines (IL-1b, IL-12p70, IFN-g, IL-6, IL-8, IL-10, and TNF) were measured by the Mouse Proinflammatory 7-Plex Ultra-Sensitive Kit (Meso Scale Discovery, Gaithersburg, MD). All of the ELISAprocedures were performed according to the manufacturers’ instructions.

PCR arrays

An RT2 profiler Type I IFN Response PCR Array (PAMM-016/PAMM-016Z; SABioscience) was performed following the manufacturer’s pro-tocol. Briefly, splenic RNA was purified from spleens as described aboveand quantitated using spectrometry. To avoid the contaminations of ge-nomic DNA, 1 mg splenic RNAwas treated with 0.5 U DNase (Invitrogen)for 15 min at room temperature and then 8 min at 70˚C to inactivate theDNase. cDNA was synthesized using the RT2 First Strand kit (SABio-science) and then loaded onto the 384-well PCR array plate for amplifi-cation on the LightCycler 480 PCR system (Roche). PAMM-016Z is anupdated version kit of PAMM-016. A list of 55 genes that are in commonbetween two versions of kit is provided in Supplemental Table I. Data wereanalyzed using the online software provided by SABioscience (http://www.sabiosciences.com/pcr/arrayanalysis.php).

Blockade of cytokines and chemokines in vivo

A dose of 0.5 mg functional-grade (azide-free, sterile-filtered, and withendotoxin levels of ,2.0 EU/mg) anti-mouse IFN-g Ab (5) (clone XMG1.2),anti-mouse IL-6 Ab (6) (clone MP5-20F3), or isotype-matched controlAbs (all from BioXCell, West Lebanon, NH) was administered i.p. toIrf-32/272/2 mice on days 21, 0, 1, and 2 postinfection. A dose of 0.2mg functional-grade anti-mouse CXCR3 Ab (16) (clone 173), anti-mouseCXCL10 Ab (17) or isotype-matched controls (BioXCell) was adminis-tered i.p. to Irf-32/272/2 mice on days21, 0, and 1 postinfection. Survivalwas monitored for 30 d after virus infection, and tissue viral load 3 d aftervirus challenge was measured by qRT-PCR as described above.

Statistical analysis

Data were analyzed by unpaired t tests or ANOVA by Dunnett multiplecomparison tests using GraphPad Prism (GraphPad Software). All errorbars represent SD. The p values , 0.05 were considered significant.

ResultsThe combined loss of IRF-3 and IRF-7 leads to enhanced viralburden in tissues

To investigate the contributions of IRF-3 and IRF-7 to the sus-ceptibility of mice to DENV infection, four different strains of mice(wild-type, Irf-32/2, Irf-72/2, and Irf-32/272/2 mice) were infectedwith 13 1011 or 13 1012 GE DENV2 strain S221, and mice weremonitored for survival. The results showed all four strains of micesurvived the infection through the monitoring period of 30 d (datanot shown), unlike mice lacking type I IFN receptor, which developa lethal disease between days 4 and 7 postinfection with 1012 GE ofS221 (10). Thus, neither IRF-3 nor IRF-7 is required for control oflethal DENV infection.To further define how the deficiency of IRF-3 and IRF-7 affects

the susceptibility of mice to DENV infection, we measured thevirus burden in mice tissues. Fig. 1 shows that increased levels ofviremia were detected from Irf-32/272/2 mice at 12, 18, 24, and72 h postinfection (hpi) relative to wild-type mice using plaqueassay (78-, 2,630-, 16,218-, and 76-fold more, respectively; p ,0.001) and qRT-PCR (509-, 45,533-, 158,489-, and 2,080-foldhigher, respectively; p , 0.001). IFNAR2/2 mice exhibited evenhigher levels of viremia across all time points. In the spleen,higher than wild-type levels of viral RNA were observed fromIrf-32/272/2 mice at all four time points postinfection (95-,1093-, 1672-, and 2706-fold higher, respectively; p , 0.001) (Fig.1B). Among wild-type and IRF single-deficient mice, Irf-72/2 mice

2 ROLES OF IRF-3 AND IRF-7 IN DENV INFECTION IN VIVO

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

have higher levels of DENV RNA in the spleen than wild-typemice (17-, 17-, and 18-fold higher at 12, 18, and 24 hpi, respec-tively, p , 0.001; 8-fold higher at 72 hpi, p , 0.05). In contrast,Irf-32/2 mice had low, wild-type levels of viral RNA in the spleen.In the livers of infected Irf-32/272/2 mice, viral RNA levels

increased significantly starting at 18 hpi (131-, 1148-, and 17-foldhigher than wild-type mice at 18, 24, and 72 h, respectively), whereasthe other three strains of mice had low, near the limit of detection,levels of virus (Fig. 1C). Similarly, higher levels of viral RNA weredetected in the kidneys of Irf-32/272/2 mice as compared withwild-type and single-deficient mice (18-, 1820-, and 260-fold higherthan wild-type mice at 18, 24, and 72 h, respectively) (Fig. 1D).We also confirmed viral replication in the spleen, the initial

target tissue of DENV in this mouse model, using immuno-histochemistry. Viral NS3 protein was readily detected at 24 hpiin the splenic marginal zones of Irf-32/272/2 mice (Fig. 1E). Inagreement with the qRT-PCR data shown in Fig. 1B, few NS3-expressing cells in Irf-72/2 and no NS3-positive cells in wild-type and Irf-32/2 spleens were detected. Similar results wereobserved at 12 hpi (data not shown). Taken together, these datademonstrate that IRF-3 and IRF-7, either individually or together,are dispensable to preventing lethal DENV infection. However,IRF-3 and IRF-7, in combination, control viral infection in mul-tiple tissues, with IRF-7 playing a slightly more important rolethan IRF-3 in restricting viral replication in the spleen.

Diminished type I IFN expression in response to DENVinfection is observed in Irf-72/2 and Irf-32/272/2 micebut not Irf-32/2 mice

As IRF-3 and IRF-7 regulate type I IFN production during viralinfections, we next examined type I IFN expression at the proteinand mRNA level in the infected mice. Serum IFN-a levels in wild-type and Irf-32/2 mice were elevated, starting at 6 hpi, peaked at

12 hpi, and lasted through 24 h, dropping back to the baselineat 72 hpi (Fig. 2A). However, a significantly diminished but notabrogated serum IFN-a response was observed in Irf-72/2 miceduring 6–24 hpi as compared with wild-type or Irf-32/2 mice (Fig.2A). Similarly, a reduced level of serum IFN-b was found in Irf-72/2 mice (Fig. 2B). These data indicate that IRF-7 contributes tothe initial stage (0–6 hpi) of type I IFN production, and in theabsence of IRF-7, an IRF-7–independent mechanism (likely IRF-3) compensates at the later stage (6–12 hpi) to achieve a partialIFN response. Compared to the single-deficient mice, neither IFN-anor IFN-b was detected above the naive mouse levels in the serumof the double-deficient mice. The dramatic loss of type I IFN se-cretion in the serum due to the combined absence of IRF-3 andIRF-7 indicates that IRF-3 and IRF-7, together, are the major reg-ulators of type I IFN production during DENV infection.To further dissect the expression profile of IFN-a/b, we next

investigated the splenic type I IFN response at the transcriptionallevel by qRT-PCR and obtained the relative IFN-a and IFN-bmRNA expression levels in infected animals over naive mice. Weobserved a significantly lower level of IFN-a mRNA in Irf-32/2

72/2 mice (0.83-fold) at 12 hpi relative to wild-type mice (2.44-fold; p , 0.01) (Fig. 2C). A gradual increase in the IFN-a mRNAexpression level was observed during 18–72 hpi. A similar ex-pression pattern of IFN-b mRNA was observed (Fig. 2D). Thesedata suggest that low levels of splenic type I IFN mRNA ex-pression are induced in a delayed manner in the absence of IRF-3and IRF-7 in mice with DENV infection.

IFN-stimulated gene induction in response to DENV infectionis delayed in Irf-32/272/2 mice

To further identify the antiviral genes that are controlled by IRF-3and IRF-7 signaling during DENV infection, we next examineda panel of IFN-stimulated gene (ISG) expression in the spleens of

FIGURE 1. Levels of virus bur-

den in DENV-infected mice. Wild-

type (WT), Irf-32/2, Irf-72/2, and

Irf-32/272/2 were infected i.v. with

1011 GE of DENV2 strain S221.

Levels of viral RNA in the serum

(IFNAR2/2mice included) (A), spleen

(B), liver (C), and kidney (D) at 12,

18, 24, and 72 hpi were determined

by qRT-PCR. Data were presented as

viral GE per milliliter of serum or per

copy of 18S rRNA of tissues. Each

time point represents 6–10mice from

two independent experiments, with

mean 6 SD from the replicates, and

the limit of detection is denoted by

a dotted line. Viral titers between the

gene-deficientmice and theWTmice

were compared by ANOVA followed

by Dunnett multiple comparisons

tests. (E) Spleens harvested from

mice at 24 hpi were sectioned, fixed,

and stained with NS3 Ab (red). Nu-

clei were counterstained with DAPI

(blue). Original magnification 35.

Images are representative of two in-

dependent experiments. *p , 0.05,

**p, 0.01, ***p, 0.001, ****p,0.0001.

The Journal of Immunology 3

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

wild-type, Irf-32/2, Irf-72/2, and Irf-32/272/2 mice at 12 or 24 hafter DENV infection using a quantitative PCR array kit. The foldchanges in expression of 55 genes from infected mice were nor-malized to that of strain-matched naive mice. At 12 hpi, 28 (50.9%),

31 (56.4%), and 31 (56.4%) of the genes were induced .3-foldover naive in wild-type, Irf-32/2, and Irf-72/2 mice, respectively,whereas only 2 (3.6%) of the 55 genes were induced .3-fold overnaive in Irf-32/272/2mice (Fig. 3A, left panel). By 24 hpi, 22 (40%),

FIGURE 2. Type I IFN expression in response to

DENV infection in mice. Wild-type (WT), Irf-32/2,

Irf-72/2, and Irf-32/272/2 mice were infected i.v.

with 1011 GE of DENV2 strain S221. (A and B)

Sera from mice were collected at 0, 6, 12, 18, 24,

and 72 hpi and analyzed for IFN-a and IFN-b by

ELISA. Each time point represents 5–10 mice from

two independent experiments, with mean 6 SD

from the replicates, and the limit of detection is

denoted by a dotted line. Serum levels of IFN-a

and IFN-b were compared by ANOVA followed

by Dunnett multiple comparisons tests. (C and D)

Total RNA was extracted from spleens of WT and

Irf-32/272/2 mice. Levels of IFN-a and IFN-b

mRNA were determined by qRT-PCR and nor-

malized to 18S rRNA of the spleens. Each time

point represents 6–10 mice from two independent

experiments, with mean 6 SD from the replicates.

Expression levels of IFN-a and IFN-b were com-

pared by unpaired t tests. *p , 0.05, **p , 0.01,

***p , 0.001, ****p , 0.0001.

FIGURE 3. Comparison of ISG induction in

DENV-infected mice. Wild-type (WT), Irf-32/2,

Irf-72/2, and Irf-32/272/2 mice were infected i.v.

with 1011 GE of DENV2 strain S221. (A) ISG ex-

pression in spleens from each strain of mice was

evaluated using RT2 profiler Type I IFN Response

PCR Array at 12 and 24 hpi. Gene expression

of transcripts with a difference of $3-fold over

naive was illustrated (black, upregulated; gray, no

change; white, downregulated). (B) Gene induction

in the spleens of WT, Irf-32/2, and, Irf-72/2 mice

was expressed as fold-increase over naive for each

strain of mice at 12 hpi. Fold change was compared

by ANOVA followed by Dunnett multiple com-

parisons tests. (C) Gene induction in the spleens of

WT and Irf-32/272/2 mice was expressed as fold-

increase over naive for each strain of mice at 12 or

24 hpi and compared by unpaired t tests. Dotted

lines denote 3-fold change over strain-matched naive

mice. Data represent mean fold-change of three

mice per strain at each time point. *p , 0.05, **p ,0.01, ***p , 0.001, ****p , 0.0001.

4 ROLES OF IRF-3 AND IRF-7 IN DENV INFECTION IN VIVO

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

16 (29.1%), 13 (23.6%), and 21 (38.2%) genes were induced morethan 3-fold over naive in wild-type, Irf-32/2, Irf-72/2, and Irf-32/2

72/2 mice, respectively (Fig. 3A, right panel). Overall, most ISGswere rapidly induced at 12 hpi, except in Irf-32/272/2 mice, whichshowed a delayed ISG expression until 24 hpi. More than 95% ofgenes induced at the later time point (24 hpi) overlapped thoseinduced at the early time point (12 hpi) in wild-type, Irf-32/2, andIrf-72/2 mice. Although Irf-72/2 mice showed similar numbers ofupregulated ISGs as wild-type and Irf-32/2 mice at 12 hpi, fivegenes (Ifna2, Ifna4, Ifnb1, Oas1a, and Oas2) were induced to alower level in Irf-72/2 mice relative to wild-type or Irf-32/2 mice(Fig. 3B), suggesting that IRF-7 plays a more important role thanIRF-3 in inducing robust expression of these particular five genesat the early stage of DENV infection.In addition, upregulation of Bst2, H2-T10, Ifna4, Isg20,Myd88,

Nmi, and Oas1b was observed in Irf-32/2 and Irf-72/2 mice, butnot in Irf-32/272/2 mice at both time points, indicating thesegenes may be regulated via an IRF-3– or IRF-7–dependent mecha-nism. The two genes, Cxcl10 (3.36-fold higher relative to naivemice) and Ifna2 (3.32-fold higher relative to naive mice) that wereslightly upregulated in Irf-32/272/2 mice at 12 hpi, represent theISGs that were induced early and independently of both IRF-3and IRF-7 in response to DENV infection. By 24 h postinfection(hpi), a total of 21 ISGs were induced in Irf-32/272/2 mice, 13 ofwhich showed even higher fold change in Irf-32/272/2 than wild-type mice (Supplemental Table I), revealing the presence of a robustISG response in the double-deficient mice at a later stage of DENVinfection through an IRF-3– and IRF-7–independent pathway. Fig.3C summarizes the fold changes in 11 genes between wild-typeand Irf-32/272/2 mice at both time points.

High level of circulating cytokines/chemokine in Irf-32/272/2

mice during the viral clearance stage

Thus far, our studies have shown that, in the combined absenceof IRF-3 and IRF-7, type I IFN and ISG expression are inducedat a later stage of DENV infection, and this delayed IFN andISG response is highly effective in mediating viral clearance andpreventing lethal disease. These findings reveal an important con-tribution of the IRF-3– and IRF-7–independent pathway in innateantiviral immunity to DENV. To start understanding the nature ofthis pathway, serum levels of innate inflammatory cytokines/che-mokines in S221-infected wild-type versus Irf-32/272/2 mice wereexamined. Although type III IFN was undetectable in the serum ofall infected mice (data not shown), Irf-32/272/2 mice had signifi-cantly higher than wild-type levels of IFN-g (1320-fold; p ,0.01), IL-6 (27.6-fold; p , 0.01), and CXCL10 (2.6-fold; p ,0.01) 24 hpi (Fig. 4A–C), when virus replication peaked andstarted to be cleared as shown in Fig. 1. Similarly, IL-8, IL-12p70, and TNF were also elevated by 7.1- (p , 0.01), 137- (p ,0.001), and 47.8-fold (p, 0.01), respectively, in Irf-32/272/2 micerelative to wild-type animals at the 24-h time point (SupplementalFig. 1A–C). By 72 hpi in Irf-32/272/2 mice, serum levels of all sixcytokines decreased, but serum IFN-g, IL-6, CXCL10, and IL-8levels remained significantly high compared with wild-type mice(p , 0.05). These data suggest that these cytokines/chemokinesmay contribute to the control of virus replication in Irf-32/272/2

mice.

IFN-g partially contributes to protection against DENVinfection in Irf-32/272/2 mice

Our published studies comparing the phenotype of DENV-infectedmice lacking type I IFN receptor alone versus mice lacking bothtype I and II IFN receptors have demonstrated a critical role forIFN-g receptor signaling in protection against DENV infection

and disease (4, 5). Therefore, we next evaluated the contributionof IFN-g to the anti-DENV immunity in Irf-32/272/2 mice. Forthis purpose, Irf-32/272/2 mice were administered neutralizingAbs against IFN-g, IFN-g–induced molecule CXCL10, CXCR3(CXCL10 receptor), or IL-6 (for comparison as a molecule notlikely to be directly involved in the IFN-g antiviral pathway),followed by challenge with 1011 GE of S221. At this viral chal-lenge dose, mice lacking type I and II IFN receptors in either 129/Sv or C57BL/6 background develop an early lethal disease 4–6 d postinfection (5, 9). A total of 100% of Irf-32/272/2 micetreated with blocking Ab against IFN-g, CXCL10, or IL-6 activitysurvived DENV challenge (data not shown). However, neutrali-zation of IFN-g activity in Irf-32/272/2 mice resulted in a sig-nificant increase in viral RNA levels in all tissues examined(serum, liver, kidney, and spleen; p , 0.001) relative to isotypecontrol Ab-treated mice. The lack of IL-6 activity, which sharesthe same isotype control Ab with IFN-g, did not lead to a differ-

FIGURE 4. High levels of circulating cytokines/chemokine in Irf-32/2

72/2 mice during the viral clearance stage. Wild-type (WT) and Irf-32/2

72/2 mice were infected i.v. with 1011 GE of DENV2 strain S221. Levels

of IFN-g (A), IL-6 (B), and CXCL10 (C) in the serum of DENV-infected

WT and Irf-32/272/2 mice at 12, 24, 18, and 72 hpi were detected by

ELISA. Each time point represents six to eight mice from two independent

experiments, with mean 6 SD from the replicates, and the limit of de-

tection is denoted by a dotted line. Data between two groups of mice were

compared by unpaired t tests. *p , 0.05, **p , 0.01.

The Journal of Immunology 5

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

ence in tissue viral RNA levels (Fig. 5A). Blockade of CXCL10activity and CXCL10 receptor CXCR3 resulted in higher viralRNA levels than isotype-treated mice in the serum (Fig. 5B).Moreover, a significant decrease in serum CXCL10 level wasobserved in IFN-g–neutralized Irf-32/272/2 mice relative to iso-type control Ab–treated Irf-32/272/2 mice (p , 0.05) (Fig. 5C),revealing that some CXCL10 was directly induced by IFN-g andinvolved in the IFN-g–mediated virus control. These data showthat IFN-g, CXCL10, CXCR3, but not IL-6, contribute to DENVrestriction in tissues of Irf-32/272/2 mice. However, neutraliza-tion of IFN-g, CXCL10, or CXCR3 activity had no impact in thesurvival phenotype of Irf-32/272/2 mice upon DENV challenge,providing further supporting an important role for the IRF-3– andIRF-7–independent mechanism in the host defense against DENV.

ISG response induced by the IRF-3 and IRF-7–independentpathway

To identify the ISGs that may be involved in the IRF-3– and IRF-7–independent signaling against DENV, ISG response was evaluatedby PCR array in IFN-g–neutralized Irf-32/272/2 mice 24 hpi(Supplemental Table II). Among the 55 genes examined, 14 geneswere upregulated .3-fold over naive in the isotype control Ab-treated Irf-32/272/2 mice, whereas only 4 genes (Cxcl10, Ifit1,Ifit3, and Mx2; p , 0.01) were upregulated in the IFN-g–neu-tralized Irf-32/272/2 mice (Fig. 6A). Fig. 6B summarizes the foldchanges of genes comparing control and IFN-g–neutralized Irf-32/272/2 mice. Thus, the 10 genes that were not activated in IFN-g–neutralized Irf-32/272/2 mice represent the ISGs induced bythe IFN-g pathway. In contrast, the four genes that were upregulated

in IFN-g–neutralized Irf-32/272/2 mice can potentially functionin the absence of IRF-3, IRF-7, and IFN-g, thereby identifyingCxcl10, Ifit1, Ifit3, andMx2 as the ISGs that can be induced via theIRF-3– and IRF-7–independent pathway during DENV infectionin mice.

DiscussionIRF-3 and IRF-7 are master transcriptional factors that regulatetype I IFN gene (IFN-a/b) induction and innate immune defenseafter virus infection. In this study, we investigated the role of IRF-3and IRF-7 for antiviral innate immunity to DENV using IRF-3– and/or IRF-7–deficient mice. The viral replication kinetics in the in-fected mice was linked to the patterns of type I IFN induction, ISGexpression, and cytokine/chemokine secretion to determine thecontributions of IRF-3 and IRF-7 to anti-DENV innate immunity.The results demonstrated that IRF-7 played an essential, nonre-dundant role in mediating the initial type I IFN response and thatIRF-3 and IRF-7 played a compensatory, redundant role in in-ducing the type I IFN response at a later phase of DENV infection.Additionally, the resistance of IFN-g–neutralized Irf-32/272/2

mice to DENV-induced lethal disease revealed a critical role forthe IRF-3– and IRF-7–independent pathway that does not involveIFN-g signaling in the host defense against DENV.Based on Irf-72/2 mice exhibiting slower viral clearance, re-

duced levels of type I IFN response, and lower transcriptionalsignals of ISGs as compared with wild-type or Irf-32/2 mice uponDENV challenge, we conclude that IRF-7 is required for mount-ing the initial type I IFN response as early as 6 h after DENVinfection. Although IRF-3 is not required for this initial type I IFN

FIGURE 5. Blockade of cytokines/chemokines

in DENV2-infected Irf-32/272/2 mice. Irf-32/272/2

mice were infected i.v. with 1011 GE of DENV2

strain S221. Viral RNA levels in serum and tissues

of Irf-32/272/2 mice, in which neutralizing Abs

against IFN-g, IL-6, or isotype control (n = 3–10)

(A) or CXCR3, CXCL10, or isotype control were

administrated (n = 6–12) (B), were determined

72 hpi. Data represent two independent experiments.

Viral RNA levels were compared by ANOVA fol-

lowed by Dunnett multiple comparisons tests. Bars

indicate geometric means. (C) Serum CXCL10

from anti–IFN-g–treated or isotype control Ab–

treated Irf-32/272/2 mice at 72 hpi was detected

by ELISA and compared by unpaired t test. The

limit of detection is denoted by a dotted line. *p,0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001.

6 ROLES OF IRF-3 AND IRF-7 IN DENV INFECTION IN VIVO

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

response, it plays an important role in inducing a robust antiviralresponse in the early phase, as the combined absence of IRF-3 andIRF-7 leads to higher viral burden in the spleen than the absenceof IRF-7 alone. At a later phase of infection, IRF-3 and IRF-7compensate the absence of each other and thus only the combinedloss of both IRF-3 and IRF-7 results in impaired viral restrictionand delayed viral clearance. In this mouse model of DENV in-fection, DENV replicates in the spleen first, followed by replica-tion in other lymphoid tissues, and finally in nonlymphoid tissues(18). Therefore, besides the temporal aspect of IRF-3 and IRF-7action, the presence of higher than wild-type or IRF-3–deficientlevels of viral RNA in the spleen of Irf-72/2 mice at both earlyand late time points postinfection suggest that IRF-7 may playa more important role in lymphoid than nonlymphoid tissuesduring the host response to DENV.Consistent with previous studies of encephalomyocarditis vi-

rus, HSV, vesicular stomatitis virus (19), and a related flavivirus,West Nile (WNV) virus (15), Irf-32/2 mice in this study showeda nearly normal type I IFN response after DENV infection. Incontrast, a deficiency of IRF-7 resulted in blunted type I IFNproduction, decreased ISG response, and increased viral burden inIrf-72/2 mice, also in agreement with the findings in WNV (20).Furthermore, the blunted systemic type I response and delayed butnot abrogated transcriptional signals of IFN-b observed in DENV-infected Irf-32/272/2 mice paralleled the findings from WNV-infected Irf-32/272/2 mice and primary macrophages (a physio-logically relevant target cells for DENV). Therefore, the earlyprotective IFN-a/b response against DENV occurs through IRF-7signaling. Based on our previous study demonstrating the MAVSpathway is responsible for the initial type I IFN response to DENV

in this mouse model (10), we conclude that the initial innate an-tiviral response to DENV is mediated by MAVS and IRF-7 sig-naling downstream of MAVS. At a later phase of infection inIrf-32/272/2 mice, the type I IFN and ISG responses in thespleen were restored, and virus replication was controlled throughthe combination of the IFN-g and IRF-3/7–independent pathways.Profiling of the ISG and chemokine/cytokine responses in

DENV-infected wild-type versus Irf-32/272/2 mice revealedCXCL10 as a key molecule that may be involved in antiviralresponses mediated by both the IFN-g and IRF-3/7–independentpathways. Indeed, neutralization of IFN-g activity in Irf-32/272/2

mice resulted in reduced but not absent levels of CXCL10,and CXCL10 or CXCR3 blockade led to increased viral burdenin Irf-32/272/2 mice. A published study using CXCL102/2 andCXCR32/2 mice has demonstrated an important role for CXCR3and CXCL10 in protection against paralytic disease induced byintracerebral DENV inoculation (21). In contrast in this study,Irf-32/272/2 mice with CXCL10 or CXCR3 blockade did notdevelop the lethal dengue hemorrhagic fever/dengue shock syn-drome–like disease that is typically observed in S221-infectedmice lacking both type I and II IFN receptors. These resultssuggest different innate immune mechanisms restrict DENV inneuronal versus nonneuronal tissues.Recent studies with other viral infection models have begun to

identify the mechanisms responsible for stimulating the innateantiviral defense independently of IRF-3 and IRF-7. IRF-1, ratherthan IRF-3 or IRF-7, was shown to be required for IFN-g–mediated inhibition of MNV replication (22). We also observedan increased level of IRF-1 in Irf-32/272/2 mice at 24 hpi, whenIFN-g starts to control DENV replication, suggesting IRF-1 may

FIGURE 6. Comparison of ISG induction in

DENV2-infected Irf-32/272/2 mice. Irf-32/272/2

mice were infected i.v. with 1011 GE of DENV2

strain S221. ISG induction in spleens from anti–

IFN-g– or isotype control Ab–treated Irf-32/272/2

mice was evaluated using RT2 profiler Type I IFN

Response PCR Array at 24 hpi. (A) Gene expres-

sion of transcripts with a difference in expression

of $3-fold over naive was illustrated (black, up-

regulated; gray no change; white, downregulated).

(B) Gene induction in the spleens anti–IFN-g– or

isotype control Ab–treated Irf-32/272/2 mice was

expressed as fold-increase over naive. Data repre-

sent mean fold-change of three mice per strain. The

dotted line denotes 3-fold change over naive mice.

Data between anti–IFN-g– or isotype control Ab-

treated Irf-32/272/2 mice were compared by un-

paired t tests. *p , 0.05, **p , 0.01.

The Journal of Immunology 7

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from

be responsible for IFN-g action against DENV. In addition, ele-vated levels of IFN-g, IL-6, and TNF were consistent with thefinding in chikungunya virus–infected Irf-32/272/2 mice at day 2(23). In the case of WNV, the IFN-b response in myeloid dendriticcells is MAVS dependent but IRF-1, IRF-3, IRF-7, IRF-8, andMyD88 independent (24). Based on the most recent publicationreporting that the MAVS-dependent induction of ISGs in WNV-infected mDCs can occur through an IRF-5–dependent pathway inthe absence of IRF-3 and IRF-7 (25), IRF-5 is a likely candidatefor the IRF-3– and IRF-7–independent signaling against DENV.Collectively, these results demonstrate that IRF-3 and IRF-7 are

redundant, albeit IRF-7 plays a more important role than IRF-3 ininducing the initial IFN-a/b response; only the combined actionsof IRF-3 and IRF-7 are necessary for efficient control of earlyDENV infection; and the late, IRF-3– and IRF-7–independentpathway contributes to anti-DENV immunity. Future studies an-alyzing the whole genome-wide expression of ISGs should furtherdefine the molecular basis of this important host defense pathwayagainst DENV and other important viruses.

DisclosuresThe authors have no financial conflicts of interest.

References1. Guzman, M. G., S. B. Halstead, H. Artsob, P. Buchy, J. Farrar, D. J. Gubler,

E. Hunsperger, A. Kroeger, H. S. Margolis, E. Martı́nez, et al. 2010. Dengue:a continuing global threat. Nat. Rev. Microbiol. 8(Suppl): S7–S16.

2. Deen, J. L., E. Harris, B. Wills, A. Balmaseda, S. N. Hammond, C. Rocha,N. M. Dung, N. T. Hung, T. T. Hien, and J. J. Farrar. 2006. The WHO dengueclassification and case definitions: time for a reassessment. Lancet 368: 170–173.

3. Julander, J. G., S. T. Perry, and S. Shresta. 2011. Important advances in the fieldof anti-dengue virus research. Antivir. Chem. Chemother. 21: 105–116.

4. Shresta, S., J. L. Kyle, H. M. Snider, M. Basavapatna, P. R. Beatty, and E. Harris.2004. Interferon-dependent immunity is essential for resistance to primarydengue virus infection in mice, whereas T- and B-cell-dependent immunity areless critical. J. Virol. 78: 2701–2710.

5. Prestwood, T. R., M. M. Morar, R. M. Zellweger, R. Miller, M. M. May,L. E. Yauch, S. M. Lada, and S. Shresta. 2012. Gamma interferon (IFN-g) re-ceptor restricts systemic dengue virus replication and prevents paralysis in IFN-a/b receptor-deficient mice. J. Virol. 86: 12561–12570.

6. Zellweger, R. M., T. R. Prestwood, and S. Shresta. 2010. Enhanced infection ofliver sinusoidal endothelial cells in a mouse model of antibody-induced severedengue disease. Cell Host Microbe 7: 128–139.

7. Shresta, S., K. L. Sharar, D. M. Prigozhin, P. R. Beatty, and E. Harris. 2006.Murine model for dengue virus-induced lethal disease with increased vascularpermeability. J. Virol. 80: 10208–10217.

8. Shresta, S., K. L. Sharar, D. M. Prigozhin, H. M. Snider, P. R. Beatty, and E. Harris.2005. Critical roles for both STAT1-dependent and STAT1-independent pathwaysin the control of primary dengue virus infection in mice. J. Immunol. 175: 3946–3954.

9. Perry, S. T., M. D. Buck, S. M. Lada, C. Schindler, and S. Shresta. 2011. STAT2mediates innate immunity to Dengue virus in the absence of STAT1 via the type Iinterferon receptor. PLoS Pathog. 7: e1001297.

10. Perry, S. T., T. R. Prestwood, S. M. Lada, C. A. Benedict, and S. Shresta. 2009.Cardif-mediated signaling controls the initial innate response to dengue virusin vivo. J. Virol. 83: 8276–8281.

11. Loo, Y. M., J. Fornek, N. Crochet, G. Bajwa, O. Perwitasari, L. Martinez-Sobrido, S. Akira, M. A. Gill, A. Garcı́a-Sastre, M. G. Katze, and M. Gale, Jr.2008. Distinct RIG-I and MDA5 signaling by RNAviruses in innate immunity. J.Virol. 82: 335–345.

12. Honda, K., and T. Taniguchi. 2006. IRFs: master regulators of signalling by Toll-like receptors and cytosolic pattern-recognition receptors. Nat. Rev. Immunol. 6:644–658.

13. Randall, R. E., and S. Goodbourn. 2008. Interferons and viruses: an interplaybetween induction, signalling, antiviral responses and virus countermeasures. J.Gen. Virol. 89: 1–47.

14. Yauch, L. E., R. M. Zellweger, M. F. Kotturi, A. Qutubuddin, J. Sidney,B. Peters, T. R. Prestwood, A. Sette, and S. Shresta. 2009. A protective role fordengue virus-specific CD8+ T cells. J. Immunol. 182: 4865–4873.

15. Daffis, S., M. A. Samuel, B. C. Keller, M. Gale, Jr., and M. S. Diamond. 2007.Cell-specific IRF-3 responses protect against West Nile virus infection byinterferon-dependent and -independent mechanisms. PLoS Pathog. 3: e106.

16. Uppaluri, R., K. C. Sheehan, L. Wang, J. D. Bui, J. J. Brotman, B. Lu, C. Gerard,W. W. Hancock, and R. D. Schreiber. 2008. Prolongation of cardiac andislet allograft survival by a blocking hamster anti-mouse CXCR3 monoclonalantibody. Transplantation 86: 137–147.

17. Khan, I. A., J. A. MacLean, F. S. Lee, L. Casciotti, E. DeHaan,J. D. Schwartzman, and A. D. Luster. 2000. IP-10 is critical for effector T celltrafficking and host survival in Toxoplasma gondii infection. Immunity 12:483–494.

18. Prestwood, T. R., M. M. May, E. M. Plummer, M. M. Morar, L. E. Yauch, andS. Shresta. 2012. Trafficking and replication patterns reveal splenic macrophagesas major targets of dengue virus in mice. J. Virol. 86: 12138–12147.

19. Honda, K., H. Yanai, H. Negishi, M. Asagiri, M. Sato, T. Mizutani, N. Shimada,Y. Ohba, A. Takaoka, N. Yoshida, and T. Taniguchi. 2005. IRF-7 is the masterregulator of type-I interferon-dependent immune responses. Nature 434: 772–777.

20. Daffis, S., M. A. Samuel, M. S. Suthar, B. C. Keller, M. Gale, Jr., andM. S. Diamond. 2008. Interferon regulatory factor IRF-7 induces the antiviralalpha interferon response and protects against lethal West Nile virus infection. J.Virol. 82: 8465–8475.

21. Hsieh, M. F., S. L. Lai, J. P. Chen, J. M. Sung, Y. L. Lin, B. A. Wu-Hsieh,C. Gerard, A. Luster, and F. Liao. 2006. Both CXCR3 and CXCL10/IFN-inducible protein 10 are required for resistance to primary infection by denguevirus. J. Immunol. 177: 1855–1863.

22. Maloney, N. S., L. B. Thackray, G. Goel, S. Hwang, E. Duan, P. Vachharajani,R. Xavier, and H. W. Virgin. 2012. Essential cell-autonomous role for interferon(IFN) regulatory factor 1 in IFN-g-mediated inhibition of norovirus replicationin macrophages. J. Virol. 86: 12655–12664.

23. Rudd, P. A., J. Wilson, J. Gardner, T. Larcher, C. Babarit, T. T. Le, I. Anraku,Y. Kumagai, Y. M. Loo, M. Gale, Jr., et al. 2012. Interferon response factors 3and 7 protect against Chikungunya virus hemorrhagic fever and shock. J. Virol.86: 9888–9898.

24. Daffis, S., M. S. Suthar, K. J. Szretter, M. Gale, Jr., and M. S. Diamond. 2009.Induction of IFN-beta and the innate antiviral response in myeloid cells occursthrough an IPS-1-dependent signal that does not require IRF-3 and IRF-7. PLoSPathog. 5: e1000607.

25. Lazear, H. M., A. Lancaster, C. Wilkins, M. S. Suthar, A. Huang, S. C. Vick,L. Clepper, L. Thackray, M. M. Brassil, H. W. Virgin, et al. 2013. IRF-3, IRF-5,and IRF-7 coordinately regulate the type I IFN response in myeloid dendriticcells downstream of MAVS signaling. PLoS Pathog. 9: e1003118.

8 ROLES OF IRF-3 AND IRF-7 IN DENV INFECTION IN VIVO

by guest on Novem

ber 13, 2017http://w

ww

.jimm

unol.org/D

ownloaded from