TLR3 Increases Disease Morbidity and Mortality from
Martha Hutchens,* Kathryn E. Luker,†Peter Sottile,†Joanne Sonstein,‡Nicholas W. Lukacs,*§
Gabriel Nu ´n ˜ez,*§¶Jeffrey L. Curtis,*‡?and Gary D. Luker2*†#
Innate immunity is required for effective control of poxvirus infections, but cellular receptors that initiate the host response to
these DNA viruses remain poorly defined. Given this information and the fact that functions of TLRs in immunity to DNA viruses
remain controversial, we investigated effects of TLR3 on pathogenesis of vaccinia virus, a prototype poxvirus. We used a recom-
binant strain Western Reserve vaccinia virus that expresses firefly luciferase to infect wild-type C57BL/6 and TLR3?/?mice
through intranasal inoculation. Bioluminescence imaging showed that TLR3?/?mice had substantially lower viral replica-
tion in the respiratory tract and diminished dissemination of virus to abdominal organs. Mice lacking TLR3 had reduced
disease morbidity, as measured by decreased weight loss and hypothermia after infection. Importantly, TLR3?/?mice also
had improved survival relative to wild-type mice. Infected TLR3?/?mice had significantly reduced lung inflammation and
recruitment of leukocytes to the lung. Mice lacking TLR3 also had lower levels of inflammatory cytokines, including IL-6,
MCP-1, and TNF-? in serum and/or bronchoalveolar lavage fluid, but levels of IFN-? did not differ between genotypes of
mice. To our knowledge, our findings show for the first time that interactions between TLR3 and vaccinia increase viral
replication and contribute to detrimental effects of the host immune response to poxviruses.
2008, 180: 483–491.
tablished as a model virus for in vivo and in vitro studies of
poxvirus biology. After eradication of smallpox, vaccinia has been
studied as a vector for gene therapy and immunization against
other pathogens. However, recent concerns about smallpox or
other diseases caused by zoonotic poxviruses, such as monkeypox,
as agents of bioterrorism have focused scientific interest on defin-
ing viral and host mechanisms that regulate pathogenesis. It is
well-established that innate immunity is essential for effective host
defense against poxvirus infections, but there are notable deficien-
cies in our knowledge of cellular receptors and pathways that ac-
tivate innate immunity to poxviruses. Identifying molecules that
initiate host immunity to vaccinia virus is expected to lead to safer
vaccines for smallpox and new strategies for therapy against in-
fection with poxviruses. The need for improved therapies is high-
lighted by recent reports of direct transmission of vaccinia from
recently vaccinated U.S. military personnel to other persons, in-
cluding a recent life-threatening case in a 2-year-old boy (1).
Moreover, it is likely that knowledge about immunity to poxvi-
The Journal of Immunology,
accinia virus, a member of the poxvirus family of en-
veloped DNA viruses, was used as the vaccine to erad-
icate smallpox as a human disease. Vaccinia now is es-
ruses will be applicable to research about other respiratory infec-
tions and immunogens.
TLR are a family of receptors that recognize specific molecular
patterns associated with many bacterial and viral pathogens. Most
TLRs localize to the plasma membrane, but those that sense nu-
cleic acids (TLR3, 7, 8, and 9) typically reside in intracellular
compartments such as endosomes. In particular, TLR3 recog-
nizes dsRNA, a structure found in the genome of some viruses
and produced as replication intermediates by vaccinia virus and
many others (2). After binding dsRNA, TLR3 signals through
the adapter protein Toll-IL-1R domain-containing adapter in-
ducing IFN-? (TRIF3; also known as TICAM1) to activate tran-
scription factors including NF-?B, AP-1, and IFN regulatory
factor 3 (IRF3). These transcription factors activate expression
of multiple inflammatory cytokines, as well as type I IFNs.
Many of the molecules regulated by TLR3, such as TNF-?,
IL-6, and type I IFNs, are known to control infection with vac-
cinia virus in vivo (3–5).
Because TLR3 recognizes dsRNA, it was assumed that this re-
ceptor would be a key determinant of protective host immunity to
viral infection. However, previous studies have produced conflict-
ing conclusions about TLR3 in host immunity to viruses. TLR3
worsens morbidity and mortality in mice infected with RNA
viruses including West Nile, phleboviruses, or influenza A (6–
8). TLR3 signaling increases levels of inflammatory cytokines,
such as IL-6 and TNF-?, that disrupt tissue barriers and exac-
erbate disease compared with TLR3?/?mice. Effects of TLR3
on morbidity and mortality may be dissociated from effects of
viral replication. For example, TLR3?/?mice had improved
survival after infection with influenza A, despite greater
*Graduate Program in Immunology,†Department of Radiology,‡Division of Pulmo-
nary and Critical Care Medicine, Department of Internal Medicine,§Department of
Pathology,¶Comprehensive Cancer Center,?Department of Veterans Affairs Health
System, and#Department of Microbiology and Immunology, University of Michigan
Medical School, Ann Arbor, MI 48109
Received for publication July 16, 2007. Accepted for publication October 22, 2007.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1This work was supported by R21AI066192, R01 HL082480, and R01 HL056309
from the National Institutes of Health (NIH), and Merit Review funds from the De-
partment of Veterans Affairs. Support for imaging experiments was provided by NIH
R24CA083099 for the University of Michigan Small Animal Imaging Resource.
2Address correspondence and reprint requests to Dr. Gary D. Luker, Center for Mo-
lecular Imaging, University of Michigan Medical School, 109 Zina Pitcher Place,
A526 BSRB, Ann Arbor, MI 48109-2200. E-mail address: firstname.lastname@example.org
3Abbreviations used in this paper: TRIF, Toll-IL-1R domain-containing adapter in-
ducing IFN-?; IRF3, IFN regulatory factor 3; WT, wild type; WR, Western Reserve;
BAL, bronchoalveolar lavage; i.n., intranasal(ly); AUC, area under the curve.
Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00
The Journal of Immunology
amounts of virus in the lung. TLR3 is not universally detrimen-
tal during infection with RNA viruses, as evidenced by a sep-
arate study showing no effect of TLR3 on pathogenesis of lym-
phocytic choriomeningitis virus, vesicular stomatitis virus, and
reovirus (9). These data emphasize the potential for TLR3 sig-
naling to modulate viral replication, local manifestations of dis-
ease, and systemic effects produced by infection with some
A limited number of studies suggest that TLR3 also may rec-
ognize DNA viruses. For example, Tabeta et al. (10) demonstrated
that TLR3?/?mice have impaired innate immunity to mouse
CMV, due primarily to reduced production of type I IFNs. These
data were not confirmed by an independent study of mouse CMV
infection in TLR3?/?mice, creating uncertainty about the rele-
vance of TLR3 for initiating protective host responses to mouse
CMV in vivo (9).
Effects of TLR3 signaling on vaccinia virus have only been
investigated in vitro. In skin biopsies from mice, TLR3 signaling
promotes secretion of a cathelicidin protein (CRAMP) that inhibits
replication of vaccinia (11). Experiments with purified peritoneal
macrophages show that vaccinia replication is ?20-fold greater in
cells isolated from mice with mutant TRIF, the only known adapter
protein for TLR3 signaling (12). These data suggest that TLR3 and
TRIF signaling are activated by vaccinia virus and function to limit
viral replication. This conclusion is also supported by the fact that
vaccinia encodes at least three different proteins that antagonize
TLR signaling pathways, including those regulated by TLR3 and
TRIF (13–15). Deletion of any one of these viral genes (A46R,
A52R, or N1L) attenuates vaccinia replication and disease morbid-
ity in mouse models, indicating that TLR signaling, including
TLR3, may be important for antiviral defense in poxvirus
In the current study, we tested the hypothesis that TLR3 is nec-
essary to limit vaccinia viral infection in vivo. Unexpectedly, we
show that deletion of TLR3 limits viral replication and reduces
morbidity and mortality from intranasal (i.n.) infection with vac-
cinia. These data suggest that TLR3 signaling contributes to the
pathogenesis of severe and fatal cases of poxvirus infections.
Materials and Methods
TLR3?/?mice backcrossed to a C57BL/6 background were originally de-
veloped by the Flavell laboratory and were bred at the University of Mich-
igan (16). Wild-type (WT) C57BL/6J control mice were obtained from The
Jackson Laboratory or Harlan Sprague Dawley. Adult male and female
mice ages 6–20 wk old were used for experiments.
We prepared stocks of Vac-FL, a recombinant Western Reserve (WR)
vaccinia virus that expresses firefly luciferase, and determined viral titers as
described previously (17, 18). To construct a recombinant virus that ex-
presses GFP fused to firefly luciferase, we first amplified firefly luciferase
from pGL3 basic (Promega) with primers 5?-AAT CAG ATC TTC CAT
CAT CAA CTT CGA GAA GCT GGG TGG TGG AGG AGA AGA CGC
CAA AAA CAT AAA GAA AGG C-3? and 5?-TTA CAA GCT TTT ACA
CGG CGA TCT TTC CGC CCT TC-3?. The PCR product was ligated to
BglII and HindIII sites in pEGFP-C1 (BD Biosciences/BD Clontech). The
GFP-luciferase fusion construct was transferred to the NheI and HindIII
sites of pRB21 (gift from B. Moss, National Institutes of Health, Bethesda,
MD). The recombinant WR vaccinia virus Vac-GFL was prepared as de-
scribed previously to insert the reporter gene without deleting any viral
genes (18, 19). Vac-GFL was plaque purified four times on Vero cells.
Vero cells were maintained as described previously (18).
All animal procedures were approved by the University of Michigan Com-
mittee on the Use and Care of Animals. We infected mice i.n. with vaccinia
virus as described previously (18). Weights and rectal temperatures (Physi-
temp Instruments) were recorded immediately before infection and daily
throughout each experiment.
Bioluminescence imaging was performed using an IVIS 200 system
(Xenogen-Caliper) on each day after infection. Imaging and data analysis
were performed as described previously (18).
Viral titers in excised organs were analyzed by serial dilution on Vero cells
as described previously (18).
To prepare lungs for histology, mice were euthanized on days 3 and 7
postinfection, and lungs were inflated with 1 ml of 10% formalin in PBS.
Lungs and tracheas were excised, preserved in 10% formalin overnight or
longer, and then transferred to 70% ethanol solution. Paraffin embedding,
sectioning, and H&E staining of fixed tissue were performed by Mc-
Clinchey Histology Laboratory.
We used immunohistochemistry to identify sites of viral replication
in the lungs, based on detection of GFP from Vac-GFL. Paraffin-em-
bedded lung sections were prepared for staining using the protocol from
the Vector Laboratories ABC staining kit. Staining was performed with
1/500 dilution of rabbit polyclonal anti-GFP Ab (Santa Cruz Biotech-
nology) and a 1/200 dilution of goat anti-rabbit secondary Ab conju-
gated to HRP.
Histologic specimens were reviewed in a blinded fashion by a board-
certified pathologist with extensive experience in animal studies (G.
Nu ´n ˜ez). To quantify foci of inflammation in lung sections, we analyzed
transverse lung sections obtained through comparable portions of the upper
and lower lobes of each lung. Numbers of inflammatory foci were observed
using a ?4 objective. Mean values for numbers of foci and SEM were
determined. Comparable counts were obtained by two independent observ-
ers (K. E. Luker, G. Nu ´n ˜ez).
Bronchoalveolar lavage (BAL)
BAL was performed as described previously (20). The first milliliter of
lavage fluid was used for analyses of total protein and cytokines. Cells
collected in all 10 ml of lavage fluid were counted on a Neubauer hemo-
cytometer, cytocentrifuged, and stained with Wright-Giemsa stain (Sigma-
Aldrich) for differential cell counts. Total protein was measured by BCA
assay (Pierce). Cytokine concentrations were measured by ELISA as de-
Blood was obtained by cardiac puncture, retro-orbital bleed, or from the
abdominal aorta of euthanized mice and collected into heparinized tubes.
Blood samples were centrifuged to separate plasma from cells. TNF-?,
IL-6, and MCP-1 concentrations in plasma and BAL fluid were determined
by ELISA performed by the University of Michigan Cancer Center Cel-
lular Immunology Core Facility. IFN-? concentrations were measured us-
ing the IFN-? ELISA kit from PBL Biomedical Laboratories according to
the manufacturer’s instructions.
Lungs were harvested on day 3 or 7 postinfection after bronchoalveolar
lavage and then disaggregated by mechanical disruption in a blender
(VWR). Cells were counted and then analyzed by flow cytometry as de-
scribed previously (21). The following mAbs obtained from BD Pharmin-
gen were used: 17A2 (anti-murine CD3, rat IgG2b), RM4-4 (anti-murine
CD4, rat IgG2b), 53-6.72 (anti-murine CD8, rat IgG2b), 1D3 (anti-murine
CD19, rat IgG2a),M1/70 (anti-murine CD11b, rat IgG2b), HL3 (anti-
murine CD11c, hamster IgG1), 2.4G2 (anti-murine CD16/CD32 Fc block,
rat IgG2b), 30-F11 (anti-murine CD45, rat IgG2b), 16-10A1 (anti-murine
CD80, hamster IgG2), M3/84 (anti-murine Mac-3, rat IgG1), and RB6-
8C5 (anti-murine Ly6G Gr-1, rat IgG2b). mAbs were primarily conju-
gated with FITC, biotin, or PE; biotinylated Abs were visualized using
streptavidin-PerCP (BD Pharmingen). In addition, the following mAbs
from eBioscience were used: 16G6 (anti-murine granzyme B, rat
484 TLR3 AND VACCINIA INFECTION
IgG2b) and ebioOMAK-D (anti-murine perforin, rat IgG2a). Isotype-
matched control mAbs (BD Pharmingen or eBioscience) were tested si-
multaneously in all experiments. Data were collected on a FACScan flow
cytometer using CellQuest software (both from BD Immunocytometry
Systems) and analyzed using FlowJo software (Tree Star). A minimum of
10,000 cells were analyzed for each sample. For all analyses, percentages
for matched isotype control Abs were subtracted from values obtained for
staining with specific Abs for individual markers.
As a positive control for granzyme and perforin staining, splenocytes
were isolated from a WT C57BL/6 mouse and cultured for 4 days in RPMI
1640 with L-glutamine, 10% FBS, 5 ?M HEPES, 2.5 ?M 2-ME, 10 U/ml
penicillin, 10 ?g/ml streptomycin, 2 mM L-glutamine, and 200 ng/ml IL-2
Data were analyzed by methods listed in the text, using Microsoft Excel
or GraphPad Prism software. Significant differences are regarded as
p ? 0.05.
TLR3?/?mice are more resistant than WT to vaccinia
TLRs are known to be key regulators of innate immune responses
to specific molecular patterns associated with pathogens, but their
functions in host responses to poxviruses remain poorly defined. In
particular, TLR3 activates NF-?B- and IRF3-dependent signaling
pathways in response to dsRNA, a molecular structure produced
during replication of viruses including vaccinia (2). Functions of
TLR3 in host defense against viruses remain controversial, and
there is little information about this receptor in immunity to DNA
viruses such as vaccinia. However, the presence of dsRNA during
replication of vaccinia virus led us to hypothesize that mice lack-
ing TLR3 would be more susceptible to vaccinia infection than
To test this hypothesis, we initially infected WT and TLR3?/?
mice intranasally (i.n.) with 1 ? 105PFU Vac-FL, a rWR vaccinia
virus that expresses firefly luciferase. We monitored viral replica-
tion and dissemination with bioluminescence imaging on each day
of infection. Weights of all animals were measured daily as a sign
of disease morbidity.
Surprisingly, TLR3?/?mice had lower viral replication than
WT in the lungs, as quantified by bioluminescence imaging and
region of interest analysis of the chest (Fig. 1A). Peak values of
bioluminescence were ?10-fold lower in TLR3?/?animals at the
peak of viral replication on day 6. Mice lacking TLR3 also had
diminished spread of virus to abdominal organs, including spleen,
liver, and inguinal lymph nodes relative to WT as determined by
imaging data (Fig. 1B). On the peak of viral replication in abdom-
inal organs, bioluminescence was ?5-fold lower in TLR3?/?an-
imals than WT. Differences in viral replication were statistically
significant as quantified by area under the curve (AUC) analysis.
AUC for photon flux in the chest was 1.96 ? 106and 1.40 ? 107
for TLR3?/?and WT mice (p ? 0.01), respectively, while the
values were 6.65 ? 105and 4.00 ? 106for AUC in the abdomen
(p ? 0.01). By comparison, levels of bioluminescence at the site
of viral inoculation in the head did not differ between genotypes
(data not shown).
Viral infection of both the brain and the upper respiratory tract
contribute to bioluminescence quantified in the head region. As the
WR strain of vaccinia virus is neurotropic, plaque assays were
performed to determine viral titers in the brain. High titers of Vac-
GSL were found in both TLR3?/?and WT brains, but there was
no significant difference between the two genotypes of mice (data
not shown). Collectively, these data show that genetic deletion of
TLR3 reduces replication of vaccinia virus in the lower respiratory
tract and systemically.
TLR3?/?mice had slightly less weight loss early after infection,
although the percent weight loss was comparable for both TLR3?/?
and WT mice after day 3 (Fig. 1C). As another quantitative measure
of disease morbidity, we measured mouse rectal temperatures.
Both genotypes of mice were hypothermic on days 5 and 6 postin-
fection (Fig. 1D), consistent with previous studies showing that
mice lose body temperature after infection with vaccinia (22). No-
tably, TLR3?/?mice maintained body temperature more effec-
tively than WT animals, showing that deletion of TLR3 protected
against this aspect of morbidity. The i.n. route of infection mimics
the natural route of smallpox infection, and depending on the dose,
may be lethal. TLR3?/?mice had a slight survival advantage at
the lethal viral inoculum used here. All WT mice died on or before
day 8 after infection. By comparison, one TLR3?/?mouse sur-
vived to day 9, and another animal recovered from infection. These
data show that deletion of TLR3 limits both viral replication and
disease morbidity after inoculation of a dose of vaccinia that is
lethal to WT mice.
To ensure that protective effects of deleting TLR3 were not
limited to an overwhelming viral challenge, we infected WT
and TLR3?/?mice with 1 ? 104PFU Vac-FL, which is 1 log
lower than the prior experiment. Beginning on day 3, viral
tant to vaccinia infection. WT (filled symbols) and
TLR3?/?(open symbols) mice (n ? 5 mice/geno-
type) were infected i.n. with 1 ? 105PFU lumines-
cent reporter vaccinia virus. A and B, Chest (A) and
abdominal (B) luminescence. Whole body biolumi-
nescence imaging was performed each day postin-
fection, and photon flux in defined regions of inter-
est was quantified by Living Image software. Data
are expressed as mean values for photon flux. Error
bars, SEM. C, Weight loss. Baseline weight was
measured the day of infection, and weights were
taken each day thereafter. Values are given as mean
percentage of initial weight. D, Body temperature.
On days 5 and 6 postinfection, core body temper-
atures of infected mice and an uninfected control
were measured with a rectal probe.
TLR3?/?C57BL/6 mice are resis-
485The Journal of Immunology
replication in the head, chest, and abdomen was substantially
reduced in TLR3?/?animals relative to WT (Fig. 2, A–D). Fig.
2A shows representative images of TLR3?/?(left) and WT
(right) mice on day 4 postinfection. By AUC analysis, total
photon flux produced by Vac-FL in these sites over the course
of infection was significantly lower in TLR3?/?mice (Table I).
Imaging data for viral replication in the lungs also were con-
firmed by plaque assays showing significantly lower amounts of
vaccinia virus in TLR3?/?mice on days 4 and 7 after infection
(Fig. 2E) (p ? 0.05). In the brain, plaque assays showed a trend
toward lower amounts of vaccinia virus in TLR3?/?mice on
days 4 and 7 postinfection, but the differences were not statis-
tically significant (Fig. 2F). Taken together, these results verify
that deletion of TLR3 reduces local and systemic replication of
TLR3?/?mice also had substantially reduced disease morbidity
after vaccinia infection as compared with WT animals. Mice lack-
ing TLR3 had reduced weight loss beginning at day 2 after infec-
tion and continuing until animals recovered on day 19 (Fig. 2G).
By AUC analysis, weight loss was significantly less in TLR3?/?
mice as compared with WT (3.13 vs 1.92) (p ? 0.01). By two-
sample equal variance t test, the difference in weight loss was
significant (p ? 0.05) on days 2–7 and 10–14 postinfection. Mice
of both genotypes continued to lose weight through day 7, which
is after the peak of viral replication on days 5 and 6. This result
suggests that host immune responses mediate ongoing disease
morbidity, at least as measured by overall weight. Notably, all
TLR3?/?mice survived infection, while 60% of WT mice died by
day 12 (Fig. 2H). The surviving mice of both genotypes had equal
serum-neutralizing Ab titers on day 28. Infection of these mice
with 1 ? 106PFU Vac-FL, a dose that is 10-fold above the lethal
inoculum in naive mice, produced no bioluminescence above
luminescent reporter vaccinia virus (n ? 10 WT and n ? 11 TLR3?/?). WT are indicated by filled symbols, and TLR3?/?are denoted with open
symbols. A, Representative bioluminescence images of TLR3?/?(left) and WT (right) mice on day 4 postinfection. Photon flux is depicted on a
pseudocolor scale with red representing the highest and blue the lowest values. B–D, Quantitative bioluminescence imaging data for head (B), chest
(C), and abdomen (D) regions of interest. Whole body bioluminescence imaging was performed each day postinfection, and photon flux in defined
regions of interest was quantified by Living Image software. Data are expressed as mean values for photon flux ? SEM. E, Lung viral titers in WT
(f) and TLR3?/?(?) on day 4 or 7 postinfection (n ? 7 WT and n ? 6 TLR3?/?). Plaque data are combined results of two experiments. F, Brain
viral titers in WT (f) and TLR3?/?(?) on day 4 or 7 postinfection (n ? 7 WT day 4 and n ? 6 TLR3?/?and WT day 7). G, Weight loss. Data
are expressed as mean percent of initial weight (n mean percent weight loss ? 10 WT or 11 TLR3?/?on day 1, decreasing to n ? 2 WT and n ?
7 TLR3?/?on day 13). ?, Significant differences between genotypes (p ? 0.05). H, Survival curve. A subset of mice (n ? 5 WT and n ? 7 TLR3?/?)
were monitored throughout the experiment until they recovered or were euthanized for 30% weight loss. Five WT and four TLR3?/?mice were
euthanized at an intermediate time points.
TLR3?/?mice are resistant to vaccinia at a dose that is approximately the LD50for WT. Mice were infected i.n. with 1 ? 104PFU
Table I. AUC analysis for photon flux from Vac-FLa
1.370 ? 107
8.858 ? 105
6.411 ? 105
7.645 ? 106
5.254 ? 105
5.491 ? 105
aMice were infected i.n. with 1 ? 104PFU Vac-FL, and viral replication was
quantified with bioluminescence imaging. Data for photon flux in head, chest, and
abdomen over the full course of the experiment were quantified by AUC analysis.
bSignificant differences between genotypes of mice (p ? 0.01).
486TLR3 AND VACCINIA INFECTION
background when animals were imaged 24 h after inoculation, and
the mice did not lose any weight upon reinfection on day 30 (data
not shown). These data are consistent with previous research dem-
onstrating a key function for neutralizing Abs in long-term pro-
tection against poxvirus infection (23). Taken together, these data
demonstrate that deletion of TLR3 limits acute replication of vac-
cinia virus, morbidity, and mortality without affecting long-term
protective immunity to repeat infection.
1 ? 104PFU Vac-FL, and leukocytes in the lung were quantified and characterized by flow cytometry. A, Number of CD45?cells at indicated times divided
by number of CD45?cells in uninfected mouse of same genotype. B, Mac-3?cells in uninfected mice (day 0) and in TLR3?/?and WT mice on days 3
and 7 postinfection. C, CD19?cells; D, Ly6G?cells. Values are expressed as a percent of total cells. Error bars denote SEM. ?, Significant differences
between genotypes (p ? 0.05). ??, p ? 0.01.
TLR3?/?mice recruit fewer leukocytes to the lung in response to vaccinia. Mice (n ? 4/genotype at each time point) were infected i.n. with
i.n. with 1 ? 104PFU Vac-FL, and CD8 T cells in the lung were analyzed by flow cytometry 7 days postinfection. A, CD8?cells, expressed as a percent
of total cells. B, Granzyme-positive cells, expressed as a percent of CD8?cells. C, Granzyme and perforin double-positive cells, expressed as a percent
of CD8?cells. Error bars denote SEM. ?, Significant differences between genotypes (p ? 0.05). Positive control, Splenocytes stimulated in vitro for 4 days
TLR3?/?and WT mice produce similar quantities of activated CTLs in response to vaccinia infection. Mice (n ? 3/genotype) were infected
487 The Journal of Immunology
Flow cytometry analysis of infiltrating cells in the lung
To establish mechanisms for detrimental effects of TLR3 on pul-
monary infection with vaccinia, we infected TLR3?/?and WT
mice with 1 ? 104PFU Vac-FL i.n., which is the inoculum that
produced differences in survival. On days 3 and 7 postinfection,
mice were euthanized to recover and analyze cells from the airway
and lung interstitium by BAL and mechanical disaggregation, re-
spectively. There were no differences between genotypes of mice
in total numbers of cells or differential cell counts performed on
BAL samples on either day. We quantified numbers of CD45?
leukocytes in the lung interstitium and analyzed types of recruited
cells by flow cytometry. In uninfected TLR3?/?and WT mice,
there were no significant differences in total numbers of cells in the
lung interstitium (data not shown). In response to infection, num-
bers of leukocytes in the lung interstitium increased in both geno-
types of mice, but the relative increases in total cell numbers were
significantly lower in TLR3?/?mice (p ? 0.05) (Fig. 3A).
TLR3?/?mice had a significantly lower percentage of Mac-3?
macrophages in the lung on day 7 (p ? 0.05) (Fig. 3B), while these
animals had a greater percentage of CD19?B cells and Ly6G?
granulocytes on day 7 after infection (p ? 0.05) (Fig. 3, C and D).
In addition, uninfected TLR3?/?mice had a significantly greater
percentage of Ly6G?granulocytes than uninfected WT mice (p ?
0.01). These data suggest that TLR3 regulates overall recruitment
of leukocytes to the lung in response to vaccinia infection, al-
though, with the exception of Ly6G?granulocytes, differences in
specific cell types were not evident until relatively late in the
course of disease.
To further analyze the immune response to vaccinia in the lung,
we quantified CD8 T cell responses in TLR3?/?and WT control
mice. We infected mice with 1 ? 104PFU Vac-FL i.n. and har-
vested lungs from these animals on day 7 after infection. Cells in
the lung interstitium were recovered by mechanical disaggregation
and analyzed by flow cytometry. Activated CTLs were identified
by Ab staining for CD8, granzyme, and perforin. Granzyme and
perforin are major components of the granules through which
CTLs destroy target cells, and these molecules are indicators of
TLR3?/?and WT control mice had comparable numbers of
CD8?T cells in the absence of infection (?1 ? 105cells), and
absolute numbers of these cells increased to a similar extent in
both genotypes after vaccinia infection (?3–4 ? 105cells). In-
fected TLR3?/?mice had a significantly higher percentage of
CD8?cells than infected WT mice on day 7 (Fig. 4A) (p ? 0.05).
However, the percent granzyme positive, activated CD8?cells did
not differ significantly between TLR3?/?and WT mice (Fig. 4B).
Additionally, both genotypes of mice had comparable percentages
of granzyme and perforin double-positive CD8?cells (Fig. 4C).
The low numbers of cells with both granzyme and perforin is con-
sistent with a previous study showing that perforin is rarely ex-
pressed by T cells analyzed directly ex vivo after vaccinia infec-
tion (24). These data show that TLR3 does not affect activation of
CTL in response to vaccinia.
To further investigate lung inflammation in TLR3?/?and WT
mice, we infected mice with 1 ? 104PFU Vac-GFL, a nonattenu-
ated recombinant vaccinia virus that expresses a GFP-firefly lucif-
erase fusion protein. We used Vac-GFL for these experiments to
enable detection of the virus based on immunohistochemistry for
GFP. On day 3 after infection, histologic sections of lungs showed
focal areas of peribronchial and perivascular inflammation in WT
mice, but no foci of inflammation were present in TLR3?/?mice.
By day 7, both TLR3?/?(Fig. 5A) and WT (Fig. 5B) lungs ex-
hibited regions of peribronchial and perivascular lymphocytic in-
filtrate on day 7, with edema and epithelial necrosis. Because vac-
cinia infection in the lung produced discrete foci of inflammation,
we were able to quantify the number of foci present in both ge-
notypes of mice. We analyzed transverse histologic sections from
comparable sites in the lower lobes of each lung. On both days 3
and 7 after infection, TLR3?/?mice exhibited a trend of fewer
foci of inflammation than WT, though the difference did not reach
statistical significance (Fig. 5C). Sites of inflammation corre-
sponded to areas of infection, as determined by immunohistochem-
istry for GFP encoded by Vac-GFL (data not shown). Overall,
these data suggest that deletion of TLR3 limits the extent of in-
flammation within the lung following infection with vaccinia virus.
TLR3?/?lung, day 7 postinfection, ??12 magnification. Inset, ?23. Arrows denote foci of inflammation. B, bronchus; V, blood vessel. B, Representative
WT lung section, day 7 postinfection, ??12 magnification. Inset, ?23. Arrows denote foci of inflammation. B, bronchus; V, blood vessel; N, region of
necrosis. C, Foci of inflammation per lung section. Foci were counted on a lung section from each mouse. Data are expressed as mean number of foci per
section (n ? 6). Error bars, SEM.
Histology of vaccinia-infected lungs. Lungs were harvested on day 3 or 7 postinfection. A, Photomicrograph of representative section of
Table II. Cytokines in BAL fluida
Day 3 Day 7
9.0 ? 2.7
0.0 ? 0.0
4.2 ? 1.2
325.7 ? 15.76 325.4 ? 22.84 308 ? 21.2 319 ? 17.2
11 ? 3.5
2.0 ? 1.0
3.5 ? 0.96
9.3 ? 3.0b
16 ? 7.7
140 ? 84
0.0 ? 0.0b
7.3 ? 2.7
120 ? 63
aCytokine levels in BAL fluid on days 3 and 7 postinfection were quantified by
bSignificant differences between genotypes of mice (p ? 0.05). Data are ex-
pressed as mean values in picograms per milliliter ? SEM.
488TLR3 AND VACCINIA INFECTION
Lung and systemic cytokines
TLR3 signaling results in activation of transcription factors includ-
ing NF-?B and IRF3. These transcription factors then increase
expression and secretion of cytokines and chemokines that regu-
late inflammation and initiate antiviral responses. To determine
effects of deleting TLR3 on cytokine responses to vaccinia, we
infected TLR3?/?and WT mice with 1 ? 104PFU Vac-GFL and
collected serum and BAL samples on days 3 and 7 after infection.
Samples were analyzed for selected NF-?B- and IRF3-dependent
cytokines by ELISA. In BAL samples, there were no significant
differences in levels of TNF-?, MCP-1, or IFN-? on day 3 (Table
II). However, TLR3?/?mice had significantly less TNF-? in BAL
fluid on day 7 than WT mice (Table II, Fig. 6A) (p ? 0.05). These
cytokines (TNF-?, IL-6, MCP-1) were undetectable or present at
comparable low levels in BAL samples from uninfected mice of
both genotypes (data not shown). In addition, IFN-? levels did not
differ between uninfected and infected mice of either genotype
(data not shown), suggesting that TLR3 did not affect activation of
IRF3 in response to vaccinia infection. We also measured total
protein in BAL samples as a determinant of lung permeability in
both genotypes of mice. Both TLR3?/?and WT mice had com-
parable levels of protein in BAL samples obtained on days 3 and
7, showing that TLR3 signaling did not alter overall permeability
of lung epithelium during vaccinia infection (data not shown).
Unlike BAL fluid, plasma levels of IL-6 and MCP-1 were sig-
nificantly lower in TLR3?/?mice on both days 3 and 7 (p ? 0.01)
(Table III) (Fig. 6, B and C). By comparison, there were no sig-
nificant differences in amounts of TNF-? or IFN-? in the plasma
of TLR3?/?or WT mice. Similar to BAL samples, no IL-6 or
TNF-? was detectable in serum of uninfected TLR3?/?or WT
mice, and low levels of MCP-1 did not differ between uninfected
TLR3?/?and WT mice. IFN-? levels also did not differ between
uninfected and infected mice (data not shown). Overall, these data
indicate that deletion of TLR3 limits secretion of selected inflam-
matory cytokines in the airway and systemically without altering lev-
els of IFN-?, a key cytokine in host defense against vaccinia (25).
Fatal infections with poxviruses, including variola major, the eti-
ologic agent of smallpox, typically are characterized by systemic
signs that resemble the clinical syndrome of septic shock. These
manifestations of severe infection, including marked dysregulation
of temperature, hypotension, and multiorgan failure, are believed
to be caused by the host inflammatory response. However, cellular
receptors and molecules that initiate this damaging and potentially
fatal immune response have not been established.
In the current study, we demonstrate that TLR3 regulates ad-
verse effects of vaccinia infection in vivo. Mice lacking TLR3
were resistant to respiratory infection with vaccinia as determined
by multiple parameters, including viral replication, morbidity, and
mortality. TLR3?/?mice had significantly fewer leukocytes re-
cruited to the lung and diminished overall lung inflammation over
3 and 7 postinfection, lungs were lavaged with 1 ml of PBS. Blood samples were collected on the same days, centrifuged, and plasma were separated from
cells. Cytokine levels were measured by ELISA. Single points indicate individual mice (WT filled, TLR3?/?open), and horizontal bars are mean values.
A, BAL TNF-?, (?, p ? 0.05). B, Plasma IL-6 (?, p ? 0.01). C, Plasma MCP-1 (?, p ? 0.01).
WT mice have significantly higher cytokine levels in BAL fluid and plasma. Mice were infected i.n. with 1 ? 104PFU Vac-GFL. On days
Table III. Cytokines in plasmaa
Day 3Day 7
1 ? 0.95
9 ? 2b
50 ? 5.6b
333 ? 46.3
0 ? 0.0
2 ? 0.6b
32 ? 1.8b
285 ? 24.8
99 ? 57
13 ? 2.4b
74 ? 8.7b
454 ? 61.3
99 ? 44
3.8 ? 0.99b
46 ? 7.5b
467 ? 33.0
aCytokine levels in serum on days 3 and 7 postinfection were quantified by
ELISA. These animals were the same as those analyzed in Table II.
bSignificant differences between genotypes of mice (p ? 0.01). Data are ex-
pressed as mean values in picograms per milliliter ? SEM.
489The Journal of Immunology
the course of infection, likely contributing to the improved out-
come in these mice. Taken together, this study is the first to show
an in vivo function for TLR3 in immunity to vaccinia, as well as
being the first report of TLR3-mediating detrimental responses to
vaccinia or any other DNA virus.
Our results for adverse effects of TLR3 were unexpected given
that vaccinia encodes multiple proteins that antagonize TLR sig-
naling, including pathways activated by TLR3 (13–15). Each of
these molecules (A46R, A52R, and N1L) interferes with distinct
components of TLR signaling, although these proteins also may
affect other intracellular pathways in addition to TLR. A46R in-
teracts with the Toll-IL-1 receptor domains of TRIF, MyD88, and
TLR4, blocking NF-?B and IRF3 activation in response to TLR3
or TLR4 signaling (13). A52R can interact with and block signals
through either the IL-1R-associated kinase 2/MyD88 adoptor-like
or the TNFR-associated factor 6 transforming growth factor-?-
activated protein kinase 1 complex, and it may disrupt TLR3-de-
pendent NF-?B induction (14). Finally, N1L blocks signals to both
NF-?B and IRF3 by interacting with the I?B kinase complex (15),
but this protein also functions as Bcl-2-like inhibitor of apoptosis
(26). Viruses lacking these antagonists of TLR signaling have
lower viral replication and produce less morbidity. In addition,
TLR signaling regulates multiple effector molecules, including
type I IFNs and other cytokines, that are needed to limit vaccinia
infection in vivo. These data suggested that TLR3 would be pro-
tective against vaccinia virus, rather than exacerbating disease se-
verity as we established.
TLR3 is not the only pattern recognition receptor capable of
signaling in response to viral infection. TLR7, 8, and 9 can elicit
cytokine secretion in response to virally derived nucleic acid li-
gands (10, 27, 28). The cytoplasmic pattern recognition receptors
retinoic acid-inducible gene I (RIG-I) and melanoma differentia-
tion-associated gene 5 (MDA-5) also do so in response to polyi-
nosinic-polycytidylic acid and RNA viruses (29). Moreover, TLR4
signals through TRIF to up-regulate IRF3-dependent genes and
has been implicated in the innate immune response to vesicular
stomatitis virus (30, 31). These redundant signaling pathways may
explain why TLR3 is dispensable for IFN-? production and pro-
tection from vaccinia infection. However, they do not explain the
reduced disease severity in TLR3?/?mice.
Similar to previous studies showing detrimental effects of TLR3
in host immunity to RNA viruses, we determined that mice lacking
TLR3 had lower levels of inflammatory cytokines in the lung and
systemic circulation. In the lung, we detected elevated levels of
TNF-? in WT mice on day 7 of infection, but levels of this cyto-
kine and other inflammatory mediators were not elevated earlier.
TLR3 signaling activates the transcription factor NF-?B, which in
turn up-regulates TNF-?. TLR3 signaling in response to vaccinia
virus may elevate levels of this cytokine beyond the period needed
to control viral replication, thereby causing tissue damage in the
lung. Because levels of TNF-? in BAL fluid are comparable be-
tween genotypes at early time points, this cytokine likely is not
critical for differences between WT and TLR3?/?mice at initial
stages of infection.
In contrast to TNF-? in BAL fluid, plasma levels of IL-6 and
MCP-1 were significantly lower in TLR3?/?mice as early as day
3 postinfection, a time point at which there were minimal differ-
ences in viral burden between genotypes of mice. TLR3?/?mice
also had significantly lower levels of these cytokines on day 7. Day
7 postinfection was a time point at which disease morbidity con-
tinued to worsen, although viral burden was decreasing in mice of
both genotypes. Because IL-6 and MCP-1 are elevated at both
early and relatively late times in acute infection, our data suggest
that these cytokines regulate TLR3-dependent immune pathology
in response to vaccinia.
Unlike our results showing increased levels of NF-?B-depen-
dent cytokines in fatal poxvirus infection, Rubins et al. (32) re-
ported that transcripts of these cytokines were reduced in cyno-
molgus macaques infected i.n. or i.v. with variola major. It is likely
that differences in experimental models and design account for
these discordant results. Rubins et al. (32) analyzed patterns of
gene expression in PBMC, rather than from the entire animal. It is
possible that the overall increase in NF-?B-dependent cytokines
that we measured is mediated by parenchymal, nonhemopoietic
cells in infected organs, rather than circulating cells in the blood.
Effects of poxvirus infection on NF-?B also may differ between
mice and nonhuman primates. Finally, combining data from respi-
ratory and i.v. inoculation of virus may be a confounding factor
because the route of infection alters the types of responding im-
mune cells and the nature of the host response. Ongoing studies in
our laboratory are focused on elucidating functions of TLR3 in
hemopoietic vs parenchymal cells during vaccinia infection.
TLR3 has been reported to have a detrimental effect in West
Nile, influenza A, and phlebovirus infections (6–8), in addition to
vaccinia. This raises the question of what advantage TLR3 confers
to the host. TLR3 is not the only TLR that has been reported to
contribute to morbidity and mortality. For example, TLR2 signal-
ing increases proinflammatory cytokine secretion, morbidity, and
mortality during infection with HSV-1, another DNA virus (33).
Similarly, TLR4 signaling mediates septic shock in response to
LPS. These observations suggest that TLR signals must strike a
delicate balance between clearing infection and causing pathologic
effects on the host. It may be that supraphysiologic doses of TLR3
ligand during vaccinia infection elicit immunopathology and per-
mit greater viral replication, while TLR3 may be beneficial at
lower doses. Another possibility is that TLR3 signaling, despite
having detrimental innate immune effects, generates a more robust
adaptive and memory response. However, our findings that WT
and TLR3?/?mice have equal percentages of activated cytotoxic
lymphocytes, equivalent neutralizing Ab levels, and comparable
resistance to vaccinia rechallenge, do not support this hypothesis.
In summary, our research establishes that TLR3 signaling is
detrimental to host defense against acute infection with vaccinia
virus, making mice more susceptible to viral replication and dis-
ease morbidity. To our knowledge, these data are the first to dem-
onstrate detrimental effects of TLR3 in host immunity to a DNA
virus. Our results imply that TLR3-dependent signaling contrib-
utes to the immunopathology associated with severe and fatal pox-
virus infections. It is possible that interrupting TLR3-mediated re-
sponses can be used therapeutically to limit the severity of disease
caused by poxviruses.
The authors have no financial conflict of interest.
1. Centers for Disease Control and Prevention. 2007. Household transmission of
vaccinia virus from contact with a military smallpox vaccinee–Illinois and Indi-
ana, 2007. MMWR Morb. Mortal Wkly. Rep. 56: 478–481.
2. Weber, F., V. Wagner, S. Rasmussen, R. Hartmann, and S. Paludan. 2006. Dou-
ble-stranded RNA is produced by positive-strand RNA viruses and DNA viruses
but not in detectable amounts by negative-strand RNA viruses. J. Virol. 80:
3. Sambhi, S., M. Kohonen-Corish, and I. Ramshaw. 1991. Local production of
tumor necrosis factor encoded by recombinant vaccinia virus is effective in con-
trolling viral replication in vivo. Proc. Natl. Acad. Sci. USA 88: 4025–4029.
4. Kopf, M., H. Baumann, G. Freer, M. Freudenberg, M. Lamars, T. Kishimoto,
R. Zinkernagel, H. Bluethmann, and G. Kohler. 1994. Impaired immune and
acute-phase responses in interleukin-6-deficient mice. Nature 368: 339–342.
490TLR3 AND VACCINIA INFECTION
5. van den Broek, M., U. Muller, S. Huang, M. Aguet, and R. Zinkernagel. 1995. Download full-text
Antiviral defense in mice lacking both ?/? and ? interferon receptors. J. Virol.
6. Wang, T., T. Town, L. Alexopoulou, J. Anderson, E. Fikrig, and R. Flavell. 2004.
Toll-like receptor 3 mediates West Nile virus entry into the brain causing lethal
encephalitis. Nat. Med. 10: 1366–1373.
7. Gowen, B., J. Hoopes, M. Wong, K. Jung, K. Isakson, L. Alexopoulou,
R. Flavell, and R. Sidwell. 2006. TLR3 deletion limits mortality and disease
severity due to Phlebovirus infection. J. Immunol. 177: 6301–6307.
8. Le Goffic, R., V. Balloy, M. Lagranderie, L. Alexopoulou, N. Escriou, R. Flavell,
M. Chignard, and M. Si-Tahar. 2006. Detrimental contribution of the Toll-like
receptor (TLR)3 to influenza A virus-induced acute pneumonia. PLoS Pathog.
9. Edelmann, K., S. Richardson-Burns, L. Alexopoulou, K. Tyler, R. Flavell, and
M. Oldstone. 2004. Does Toll-like receptor 3 play a biological role in virus
infections? Virology 322: 231–238.
10. Tabeta, K., P. Goergel, E. Janssen, X. Du, K. Hoebe, K. Crozat, S. Mudd,
L. Shamel, S. Sovath, J. Goode, et al. 2004. Toll-like receptors 9 and 3 as es-
sential components of innate immune defense against mouse cytomegalovirus
infection. Proc. Natl. Acad. Sci. USA 101: 3516–3521.
11. Howell, M., R. Gallo, M. Boquniewicz, J. Jones, C. Wong, J. Streib, and
D. Leung. 2006. Cytokine milieu of atopic dermatitis skin subverts the innate
immune response to vaccinia virus. Immunity 24: 341–348.
12. Hoebe, K., X. Du, P. Gerogel, E. Janssen, K. Tabeta, S. Kim, J. Goode, P. Lin,
N. Mann, S. Mudd, et al. 2003. Identification of Lps2 as a key transducer of
MyD88-independent TIR signaling. Nature 424: 743–748.
13. Stack, J., I. Haga, M. Schroder, N. Bartlett, G. Maloney, P. Reading,
K. Fitzgerald, G. Smith, and A. Bowie. 2005. Vaccinia virus protein A46R targets
multiple Toll-like-interleukin-1 receptor adaptors and contributes to virulence.
J. Exp. Med. 201: 1007–1018.
14. Harte, M., I. Haga, G. Maloney, P. Gray, P. Reading, N. Bartlett, G. Smith,
A. Bowie, and L. O’Neill. 2003. The poxvirus protein A52R targets Toll-like
receptor signaling complexes to suppress host defense. J. Exp. Med. 197:
15. DiPerna, G., J. Stack, A. Bowie, A. Boyd, G. Kotwal, Z. Zhang, S. Arvikar,
E. Latz, K. Fitzgerald, and W. Marshall. 2004. Poxvirus protein N1L targets the
I-?B kinase complex, inhibits signaling to NF-?B by the tumor necrosis factor
superfamily of receptors, and inhibits NF-?B and IRF3 signaling by Toll-like
receptors. J. Biol. Chem. 279: 36570–36578.
16. Alexopoulou, L., A. Holt, R. Medzhitov, and R. Flavell. 2001. Recognition of
double-stranded RNA and activation of NF-?B by Toll-like receptor 3. Nature
17. Earl, P., N. Cooper, L. Wyatt, and B. Moss. 1998. Preparation of cell cultures and
vaccinia virus stocks. In Current Protocols in Molecular Biology. F. M. Ausubel,
R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl,
eds. John Wiley & Sons, Hoboken, NJ. pp. 16.16.1–16.16.11.
18. Luker, K., M. Hutchens, T. Schultz, A. Pekosz, and G. Luker. 2005. Biolumi-
nescence imaging of vaccinia virus: effects of interferon on viral replication and
spread. Virology 341: 284–300.
19. Blasco, R., and B. Moss. 1995. Selection of recombinant vaccinia viruses on the
basis of plaque formation. Gene 158: 157–162.
20. Rivera, R., M. Hutchens, K. E. Luker, J. Sonstein, J. L. Curtis, and G. D. Luker.
2007. Murine alveolar macrophages limit replication of vaccinia virus. Virology
21. Osterholzer, J., T. Ames, T. Polak, J. Sonstein, B. Moore, S. Chensue, G. Toews,
and J. Curtis. 2005. CCR2 and CCR6, but not endothelial selectins, mediate the
accumulation of immature dendritic cells within the lungs of mice in response to
particulate antigen. J. Immunol. 175: 874–883.
22. Alcami, A., and G. Smith. 1996. A mechanism for the inhibition of fever by a
virus. Proc. Natl. Acad. Sci. USA 93: 11029–11034.
23. Edghill-Smith, Y., H. Golding, J. Manischewitz, L. King, D. Scott, M. Bray,
A. Nalca, J. Hooper, C. Whitehouse, J. Schmitz, et al. 2005. Smallpox vaccine-
induced antibodies are necessary and sufficient for protection against monkeypox
virus. Nat. Med. 11: 740–747.
24. Rock, M., S. Yoder, P. Wright, T. Talbot, K. Edwards, and J. J. Crowe. 2005.
Differential regulation of granzyme and perforin in effector and memory T cells
following smallpox immunization. J. Immunol. 174: 3757–3764.
25. Deonarain, R., A. Alcami, M. Alexiou, M. Dallman, D. Gewert, and A. Porter.
2000. Impaired antiviral response and ?/? interferon induction in mice lacking ?
interferon. J. Virol. 74: 3404–3409.
26. Cooray, S., M. Bahar, N. Abrescia, C. McVey, N. Bartlett, R. Chen, D. Stuart,
J. Grimes, and G. Smith. 2007. Functional and structural studies of the vaccinia
virus virulence factor N1 reveal a Bcl-2-like anti-apoptotic protein. J. Gen. Virol.
27. Heil, F., H. Hemmi, H. Hochreim, F. Ampenberger, C. Kirschning, S. Akira,
G. Lipford, H. Wagner, and S. Bauer. 2004. Species-specific recognition of sin-
gle-stranded RNA via Toll-like receptor 7 and 8. Science 303: 1526–1529.
28. Diebold, S. S., T. Kaisho, H. Hemmi, S. Akira, and C. Reis e Sousa. 2004. Innate
antiviral responses by means of TLR7-mediated recognition of single-stranded
RNA. Science 303: 1529–1531.
29. Gitlin, L., W. Barchet, S. Gilfillan, M. Cella, B. Beutler, R. Flavell, M. Diamond,
and M. Colonna. 2006. Essential role of mda-5 in type I IFN responses to polyri-
boinosinic:polyribocytidylic acid and encephalomyocarditis picornavirus. Proc.
Natl. Acad. Sci. USA 103: 8459–8464.
30. Doyle, S., S. Vaidya, R. O’Connell, H. Dadgostar, P. Dempsey, T. Wu, G. Rao,
R. Sun, M. Haberland, R. Modlin, and G. Cheng. 2002. IRF3 mediates a TLR3/
TLR4-specific antiviral gene program. Immunity 17: 251–263.
31. Georgel, P., Z. Jiang, S. Kunz, E. Janssen, J. Mols, K. Hoebe, S. Bahram,
M. Oldstone, and B. Beutler. 2007. Vesicular stomatitis virus glycoprotein G
activates a specific antiviral Toll-like receptor 4-dependent pathway. Virology
32. Rubins, K., L. Hensley, P. Jahrling, A. Whitney, T. Geisbert, J. Huggins,
A. Owen, J. Leduc, P. Brown, and D. Relman. 2004. The host response to small-
pox: analysis of the gene expression program in peripheral blood cells in a non-
human primate model. Proc. Natl. Acad. Sci. USA 101: 15190–15195.
33. Kurt-Jones, E., M. Chan, S. Zhou, J. Wang, G. Reed, R. Bronson, M. Arnold,
D. Knipe, and R. Finberg. 2004. Herpes simplex virus 1 interaction with Toll-like
receptor 2 contributes to lethal encephalitis. Proc. Natl. Acad. Sci. USA 101:
491 The Journal of Immunology