The Journal of Immunology
Flagellin Promotes Myeloid Differentiation Factor
88-Dependent Development of Th2-Type Response1
Arnaud Didierlaurent,*† Isabel Ferrero,‡ Luc A. Otten,† Bertrand Dubois,§
Monique Reinhardt,* Harald Carlsen,¶ Rune Blomhoff,¶ Shikuo Akira,储
Jean-Pierre Kraehenbuhl,*† and Jean-Claude Sirard2*
Activation of dendritic cells (DC) by microbial products via Toll-like receptors (TLR) is instrumental in the induction of immunity.
In particular, TLR signaling plays a major role in the instruction of Th1 responses. The development of Th2 responses has been
proposed to be independent of the adapter molecule myeloid differentiation factor 88 (MyD88) involved in signal transduction by
TLRs. In this study we show that flagellin, the bacterial stimulus for TLR5, drives MyD88-dependent Th2-type immunity in mice.
Flagellin promotes the secretion of IL-4 and IL-13 by Ag-specific CD4ⴙ T cells as well as IgG1 responses. The Th2-biased
responses are associated with the maturation of DCs, which are shown to express TLR5. Flagellin-mediated DC activation requires
MyD88 and induces NF-␬B-dependent transcription and the production of low levels of proinflammatory cytokines. In addition,
the flagellin-specific response is characterized by the lack of secretion of the Th1-promoting cytokine IL-12 p70. In conclusion, this
study suggests that flagellin and, more generally, TLR ligands can control Th2 responses in a MyD88-dependent manner. The
Journal of Immunology, 2004, 172: 6922– 6930.
D
endritic cells (DC)3 play a key role in immune surveillance and have the unique capacity among APCs to stimulate naive T cells and differentiation of Th1- and Th2type CD4⫹ T lymphocytes (1–3). DC activation is controlled by
conserved signature molecules of microbes, called pathogen-associated molecular patterns (PAMP) (4). PAMP interaction with pattern recognition receptors (PRR), such as members of the Toll-like
receptor (TLR) family, triggers the signal transduction required for
coordinated activation of innate and adaptive immunities. In particular, recognition of PAMPs by PRRs can configure DCs to instruct the most appropriate Th cell response for protection (5).
TLRs (TLR1 to TLR10) share common characteristics in the
signal transduction machinery, including the myeloid differentiation factor 88 (MyD88) adapter molecule and the NF-␬B transcriptional activator (4). MyD88-deficient mice develop a Th2-biased
response after immunization with adjuvants containing TLR acti*Swiss Institute for Experimental Cancer Research and †Institute of Biochemistry,
University of Lausanne, Epalinges, Switzerland; ‡Ludwig Institute for Cancer Research, Epalinges, Switzerland; §Institut National de la Santé et de la Recherche
Médicale, Unité 404, Lyon, France; ¶Institute for Nutrition Research, University of
Oslo, Oslo, Norway; and 储Department of Host Defense, Osaka University, Osaka,
Japan
Received for publication December 5, 2003. Accepted for publication March
23, 2004.
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.
1
This work was supported by grants from Institut National de la Santé et de la
Recherche Médicale (Program AVENIR R02344ES), the Région Nord Pas de Calais
(ARCir émergence), the Fondation pour la Recherche Médicale (to J.C.S.), the Swiss
National Science Foundation (3100-067926.02 and PRN Molecular Oncology; to
J.P.K.), the European Union (QLRT-PL1999-01321 and OFES 99.041-1 and -2; to
J.P.K. and J.C.S.), and Grant 31-59165-99 (to L.A.O.).
2
Address correspondence and reprint requests to Dr. Jean-Claude Sirard, Institut de
Biologie de Lille, Groupe AVENIR, Institut National de la Santé et de la Recherche
Médicale, d’Immunité Anti-Microbienne des Muqueuses, E0364, 1 rue du Professeur
Calmette, BP 447, 59021 Lille Cedex, France. E-mail address: [email protected]
3
Abbreviations used in this paper: DC, dendritic cell; BM, bone marrow; BMDC,
BM-derived DC; MyD88, myeloid differentiation factor 88; HEL, hen egg lysozyme;
PAMP, pathogen-associated molecular pattern; PRR, pattern recognition receptor;
RPMI-5, RPMI 1640 containing 5% FCS; TLR, Toll-like receptor.
Copyright © 2004 by The American Association of Immunologists, Inc.
vators, such as CFA or the soluble Toxoplasma Ag (6 – 8). Thus,
MyD88-dependent TLR signaling is probably dedicated to the induction of Th1-biased responses (5). In contrast, Th2 commitment
is assumed to depend on DC activation via PRR unrelated to TLR,
such as Schistosoma egg Ags. However, recent studies propose
that TLR signaling contributes to the induction of Th2-type differentiation by control of DC activation (9 –13). In particular, Kaisho et al. (11) showed that MyD88-independent rather that TLRindependent DC activation seems to be involved in Th2
development.
Monomeric flagellin is a PAMP that mediates signal transduction in mammals via TLR5 and MyD88 (14, 15). Flagellin is the
structural protein subunit of the flagellum, the motility apparatus of
bacteria. Flagellin is known to elicit, in the absence of any adjuvant, a strong and sustained Ab response, which is a hallmark of
Th2-type immune responses (for review, see Ref. 16). These pioneer studies of flagellin were particularly instrumental to establish
the paradigm of Th1/Th2 differentiation. More recent work in mice
has reported that flagellin can potentate the priming of Th1-type
CD4⫹ T cells specific for the OVA Th peptide, OVA323–339 (17).
Flagellin was also shown to instruct human DCs to stimulate Th1
responses in an in vitro model (13). The development of Th1 immunity was correlated with the secretion of IL-12 p70 by activated
DC. However, these data were not supported by the study by
Means et al. (18), which showed that flagellin-activated human
DCs do not produce IL-12 p70. Similarly, activation of mouse DC
by flagellin was not observed in all conditions (17, 18). In contrast,
the potency of flagellin as an adjuvant has been established in rats
and mice (16). In this study we assessed both in vitro and in vivo
the effects of flagellin on DC maturation and T cell differentiation
using conventional mice and mice deficient for MyD88. We show
that flagellin from Salmonella typhimurium induces MyD88-dependent development of Th2 responses. Although flagellin induces
NF-␬B-dependent transcription and production of IL-12 p40,
flagellin-mediated DC activation is not associated with the secretion of the Th1-polarizing cytokine IL-12 p70 or high levels of
proinflammatory cytokines. Our observations suggest that flagellin
0022-1767/04/$02.00
The Journal of Immunology
has a specific, MyD88-dependent role in adaptive immunity by
activating DC to instruct the development of Th2-biased
responses.
Materials and Methods
Flagellin purification and reagents
S. typhimurium produces two flagellins: FliC and FljB. Isogenic FliC-producing strain SIN22 ( fljB5001::MudJ) and flagellin-deficient SIN41
( fliC5050::MudJ fljB5001::Mud-Cam) were obtained by phage transduction using strains TH714 and TH2795 as donors and ATCC14028 (American Type Culture Collection, Manassas, VA) as recipient (19). Flagellin
FliC was prepared from SIN22. Briefly, flagella were sheared from surface,
concentrated by ultracentrifugation, and heated for 30 min at 65°C to release FliC monomers (5 mg/l culture). FliC was depleted of endotoxin using
Detoxi-Gel affinity columns (Pierce, Rockford, IL). Protein concentration was
determined by the Bradford microassay (Bio-Rad, Hercules, CA), and endotoxin levels (⬍20 pg/␮g FliC) were determined by Limulus pyrochrome assay
(Associates of Cape Cod, Falmouth, MA). Similar preparations from SIN41
without endotoxin depletion were made. When specified, FliC was digested at
37°C for 30 min with 0.05% trypsin/0.02% EDTA (Biochrom, Berlin, Germany), followed by 1-h inactivation at 70°C. OVA (Sigma-Aldrich, St. Louis,
MO) and hen egg lysozyme (HEL; Appligen, Strasbourg, France) were detoxified (⬍20 pg endotoxin/␮g protein). S. typhimurium LPS (L-6511) or
phosphorothioate CpG oligonucleotide (TCCATGACGTTCCTGATGCT)
were obtained from Sigma-Aldrich and Eurogentec (Brussels, Belgium).
Mice
Six- to 16-wk-old mice (BALB/c (H-2d), C57BL/6 (H-2b), C3H/HeJ (H-2k),
Naval Medical Research Institute (NMRI) outbred mice, DO11.10 SCID
transgenic for OVA-specific TCR, MyD88⫺/⫺ on C57BL/6 background, and
(C57BL/6J ⫻ CBA/J)F1 transgenic for a transcriptional fusion between
promoter containing NF-␬B sites from the Ig␬ L chain and the firefly
luciferase (3⫻-␬B-luc)) were purchased from Harlan (Indianapolis, IN) or
were bred in our animal facility (6, 20).
DC preparation
Cell cultures were performed in RPMI 1640 containing 5% FCS (RPMI-5)
or in IMDM containing 10% FCS (IMDM-10) and supplements (Life
Technologies). Bone marrow (BM) DC precursors were obtained from
femoral and tibial bones (21). Briefly, BM cells were plated at 2 ⫻ 105
cells/ml with 10 ng/ml GM-CSF (BioSource International, Camarillo, CA)
in IMDM-10 and supplemented 3 days later with GM-CSF-containing medium. On day 6, BM-derived DCs (BMDCs) were stimulated for 18 h and
analyzed. Splenic DCs were prepared as previously described (22). Briefly,
spleens were incubated with 0.5 mg/ml collagenase D and 40 ␮g/ml DNase
I (Roche, Indianapolis, IN) in RPMI-5 for 10 min at 37°C. After mechanical dissociation, splenic DCs were enriched by centrifugation in cold isoosmotic Optiprep (Nycomed, Oslo, Norway).
Immunizations
OVA-specific CD4⫹ T cells were isolated from spleen and lymph nodes of
DO11.10 mice using MACS CD4 beads (Miltenyi Biotech, Auburn, CA)
with a purity ⬎98%. The cells were stained with 5 ␮M CFSE (Molecular
Probes, Eugene, OR), and 4 –5 ⫻ 106 cells were injected i.v. into recipient
BALB/c mice. One day later, mice were injected i.v. with OVA with or
without various stimuli. CD4⫹ V␤8⫹ splenocytes or lymph node cells were
analyzed 72 h later. OVA-specific T cell responses were analyzed in conventional mice immunized s.c. Ten days later, 2–5 ⫻ 106 lymph node cells
mixed with 2–5 ⫻ 106 irradiated splenocytes were incubated for 72 h with
RPMI-5 with or without OVA to measure the secretion of IL-4, IL-13, and
IFN-␥. Animals were immunized s.c. once or twice on days 0 and 21 with
FliC and/or OVA, and the systemic Ab response was analyzed on day 28.
Luciferase assays
BMDC or CD11c⫹ cells sorted from BMDC derived from 3⫻-␬B-luc mice
were stimulated for 30 min to 3 h. Luciferase activity was measured in cell
extracts using the luciferase assay system (Promega, Madison, WI).
HeLa S3 cells were transfected with ELAM-luc with or without pEF6/V5TOPO-hTLR5 (1 ␮g of each plasmid) using Lipofectin (Invitrogen, Carlsbad, CA) and incubated for 6 h with 100 ng/ml flagellin, and cell lysates
were tested in luciferase assays.
Abs and flow cytometry
The following mAbs conjugated to FITC, PE, CyChrome, or biotin were
used: CD11c (HL3), MHC class II (HB11.54.3), B220 (RA3.6B2), CD8␣
6923
(53.6.7), CD4 (LT4), CD80 (16-10A1), CD86 (GL-1), CD40 (3/23 and
FGK-45), CD62L (MEL-14), CD44 (IM-7), F4/80 (F4/80), HEL46 – 61
loaded in I-Ak (C4H3), and mAb isotype controls (from BD PharMingen,
San Diego, CA; or homemade) (23). Biotinylated Ab were revealed with
streptavidin conjugated to CyChrome (BD PharMingen) or allophycocyanin (Molecular Probes). For all staining, FcR were blocked by incubating
cells with mAb anti-CD32 (2.4G2). Flow cytometry was performed using
FACSCalibur and CellQuest, and sorting was performed using FACStar
(BD Biosciences, Mountain View, CA).
Quantitative RT-PCR
Total RNA was treated with DNase I (Qiagen, Chatsworth, CA) and reverse
transcribed using Superscript II (Invitrogen). Quantitative PCR was then
performed with an ABI 5700 thermocycler (Applied Biosystems, Foster City,
CA) using the SYBR Green PCR assay and the primers specific for TLR5
(CGCACGGCTTTATCTTCTCC and GGCAAGGTTCAGCATCTTCAA)
and18SribosomalRNA(TTGGCAAATGCTTTCGCTTandCGCCGCTAGA
GGTGAAATTC). Relative mRNA levels were determined using 18S RNA as
a reference. The expression of IL-4 and IFN-␥ was measured using a Light
Cycler (Roche) and the following primers: Porphobilinogen deaminase
(GAAG GATG TGCC TACC ATAC TACC TC and GGTT TTCC CGTT
TGCA GATG), IFN-␥ (GGATGCATTCATGAGTATTG and CTTTTC
CGCTTCCTGAGG), and IL-4 (CGGAGATGGATGTGCCAAAC and AGC
CCTACAGACGAGCTCACTC). Amplification plots were analyzed using the
second derivative method with LC data analysis 3.5 software (Roche).
ELISA
Cytokines were quantified by ELISA kits (IFN-␥, IL-12 p40, IL-12 p70,
and TNF-␣ from BD PharMingen; IL-4, IL-6, and IL-13 from R&D Systems, Minneapolis, MN). IL-4 secretion was also checked by bioassays
using the CTL44 cell line (24). For detection of FliC- and OVA-specific
Ab, sera were analyzed by ELISA using microplates (Maxisorp; Nunc,
Naperville, IL) coated with 100 ng of FliC or 1 ␮g of OVA/well in PBS.
Detection of Ab was performed using peroxidase-conjugated goat antimouse IgG (Bio-Rad), and titers were expressed as the reciprocal of the
highest dilution that yielded an OD492 of 0.1. Quantification of IgG1 and
IgG2a titers was performed using biotin-conjugated goat anti-mouse IgG1
and IgG2a (Caltag, San Francisco, CA). Student’s t test was used to determine statistics.
Results
Flagellin triggers DC activation
The regulation of adaptive immunity by TLR agonists is dependent
on stimulation of DCs (4). TLRs are expressed by DCs, and their
contribution to DC activation is established for TLR2, -3, -4, -6, -7,
and -9 (4). The expression of TLR5, the flagellin-specific TLR,
was previously detected in spleen or splenic macrophages (18, 25,
26). To determine whether mouse splenic DCs can respond to
flagellin as previously suggested (17), the expression of TLR5 was
assessed by real-time RT-PCR and normalized to TLR5 expression
in total splenocytes (Fig. 1a). The specificity of the assay was
checked by analysis of melting curves of various samples and by
sequencing of the PCR product. TLR5 transcript levels in splenic
DCs were ⬃10-fold higher than those in whole splenic cells. We
confirmed that TLR5 transcripts are found in splenic macrophages
at a lower level than in splenic DCs, but not at all in B lymphocytes
(data not shown). These data identified DCs as the main in vivo
TLR5-expressing splenic APCs and are in agreement with recent
observations (27). The expression of TLR5 in BMDCs was consistently ⬃10-fold less than that in splenic DCs (i.e., 0.6 –1.6 relative expression level compared with total splenocytes; Fig. 1a). In
conclusion, mouse DCs express TLR5 and are therefore equipped
to respond to flagellin.
A major concern in the analysis of bacterial product activity on
mammalian cells, in particular on DC, is contamination by a low
concentration of stimulatory signals, such as LPS or lipopeptides.
To investigate the effect of flagellin on DCs, monomeric flagellin
FliC was purified from S. typhimurium, and endotoxin contaminants were removed. Coomassie Blue staining and immunoblot
revealed a unique 52-kDa band corresponding to FliC (Fig. 1b). To
6924
FLAGELLIN/MyD88-SPECIFIC STIMULATION OF Th2-BIASED RESPONSES
FIGURE 1. Flagellin-mediated activation of DC. a, Splenic DCs from naive mice were isolated by FACS sorting as F4/80⫺/low CD11c⫹ cells population.
BMDC were isolated by CD11c-specific MACS sorting. TLR5 RNA levels were quantified by real-time RT-PCR, normalized with respect to 18S RNA
and presented as the fold increase compared with values observed in total splenocytes. b, Flagellin preparations from S. typhimurium (2 ␮g or equivalent/
lane) were analyzed on SDS-PAGE, followed by gel staining with Coomassie Blue (left) or immunoblotting using FliC-specific Ab (right). Lane 1, Flagellin
FliC; lane 2, preparation from flagellin-deficient strain; lane 3, trypsin-treated FliC. Preparation purity was assayed by immunoblot with FliC-specific serum
generated in C57BL/6 and BALB/c mice immunized with 40 ␮g FliC and CFA, followed by 20 ␮g FliC and IFA. Coomassie Blue staining and immunoblot
revealed a unique 52-kDa band corresponding to FliC. c, The effect of FliC on DCs was investigated in C3H/HeJ mice injected i.v. with PBS (dotted line),
a preparation of flagellin-deficient S. typhimurium (thin line; equivalent to 20 ␮g of FliC), or 1 ␮g of FliC (bold line). Six hours later splenic DCs were
analyzed by flow cytometry for the indicated activation markers. Histograms represent data for F4/80⫺/low CD11c⫹ gated cells. Similar results were
obtained in C57BL/6 mice. d, C3H/HeJ BMDCs were incubated for 20 h with medium alone (dotted line), 1 ␮g/ml FliC (bold line), or trypsin-treated FliC
(thin line; equivalent to 1 ␮g/ml FliC). Diagrams represent flow cytometric analysis of CD11c⫹ cells. Results are representative of five experiments. The
gray diagrams correspond to staining with an isotype control antibody. e, BMDCs from C3H/HeJ mice were incubated for 20 h with 0.5 mg/ml HEL, 1
␮g/ml FliC, and trypsin-treated FliC (equivalent to 1 ␮g/ml), alone or in combination. The median fluorescence of CD11c⫹ cells was measured by staining
with C4H3 and MHC II-specific Abs (left). C4H3-specific median fluorescence was normalized to MHC II-specific fluorescence and was expressed as a
percentage of the ratio for mock-treated DCs (right). The results shown are representative of two experiments.
The Journal of Immunology
6925
rule out any activity of protein and nonprotein contaminants, two
controls were included: trypsin-treated detoxified FliC and preparation from flagellin-deficient S. typhimurium (Fig. 1b). We first
verified that FliC preparations were able to activate mammalian
cells responding exclusively to flagellin. All flagellin FliC preparations used in the study, in contrast to controls, up-regulated the
transcription of IL-8 and CCL20 genes in epithelial intestinal cell
lines as described previously (data not shown) (19, 28, 29). Similarly, NF-␬B luciferase fusion was up-regulated in HeLa cells
transfected with human TLR5 expression plasmid upon incubation
with flagellin (data not shown). Therefore, our preparations induce
signaling in mammalian cells via flagellin, the TLR5-stimulating
ligand. We then investigated DC activation in vivo 6 h after i.v.
injection of mice with flagellin (Fig. 1c). Flagellin enhanced the
surface expression of the costimulatory molecules CD40, CD80,
and CD86 as well as the expression of MHC class II in splenic
DCs. Activation was effective at FliC doses as low as 100 ng (i.e.,
2 pmol), but was not observed after injection of trypsin-treated
FliC or preparation from flagellin-deficient S. typhimurium (Fig. 1c
and data not shown). A flagellin-specific effect on splenic DCs was
observed in LPS-hyporesponsive (C3H/HeJ), TLR2⫺/⫺, or conventional inbred animals (C57BL6 and BALB/c; Figs. 1c and 2a)
(A. Didierlaurent and T. Roger, unpublished observations). In addition, splenic B lymphocytes did not proliferate or induce MHC
class II expression upon incubation with flagellin, in contrast to
other TLR activators such as peptidoglycan (TLR2 and TLR6),
poly(I-C) (TLR3), LPS (TLR4), or unmethylated CpG DNA
(TLR9) (A. Didierlaurent, K. Burns, and B. Brissoni, unpublished
observations). Thus, our findings indicate that the stimulatory activity of flagellin preparations is independent of contaminants signaling through TLR2, TLR3, TLR4, TLR6, and TLR9. These observations also show that responsiveness to flagellin correlates
with TLR5 expression at the transcriptional level. Together, our
data show that splenic DCs, a major source of TLR5 in the spleen,
are specifically stimulated by flagellin.
To determine whether flagellin was able to induce DC maturation in the absence of other cell types, DCs were derived from BM
precursors. As observed for splenic DCs, flagellin induced the upregulation of MHC class II and costimulatory molecules on BMDCs
(Fig. 1d). Trypsin-treated FliC had no effect, and a dose-response
analysis showed that flagellin was active at concentrations as low as
10 ng/ml (data not shown). These experiments demonstrate that
flagellin induces a direct activation of DC. zsn increase in Ag processing and presentation functions is also an important feature of
DC activation. The capacity of flagellin to potentate processing
and presentation of whole HEL was evaluated using C4H3 mAb
specific for I-Ak loaded with HEL46 – 61 peptide (Fig. 1e) (23). As
described previously, HEL alone was presented by BMDCs, in
contrast to mock-treated cells. Incubation of BMDCs with HEL
and FliC resulted in 2-fold enhancement of HEL presentation. Together these data indicate that flagellin induces the functional maturation of DC.
Maturation of DC by flagellin requires MyD88
To provide evidence that TLR5 may participate in flagellin effect,
experiments were conducted in MyD88-deficient mice (6, 14).
LPS was used as a control for MyD88-independent DC activation
(30). Splenic DCs and BMDCs from MyD88⫺/⫺ mice were not
activated by flagellin or CpG in contrast to LPS (Fig. 2, a and b).
Flagellin-mediated signaling in DC is therefore strictly dependent
on the adapter molecule MyD88. TLR signaling pathways are invariably associated with transcriptional activation of NF-␬B-dependent genes. BMDC derived from transgenic mice harboring a
NF-␬B-responsive luciferase fusion were incubated with LPS,
FIGURE 2. MyD88-dependent activation of DC. a, The activation of
splenic DCs was characterized in MyD88⫺/⫺ and C57BL/6 mice injected
i.v. with PBS (dotted line) or 1 ␮g of FliC or LPS (bold line). Six hours
later, splenic DCs were analyzed by flow cytometry for the indicated markers. Histograms represent data for F4/80⫺/low CD11c⫹ gated cells. b, BMDCs derived from MyD88⫺/⫺ or C57BL/6 were incubated for 20 h with
medium alone (dotted line), FliC (1 ␮g/ml), LPS (1 ␮g/ml), or CpG (10
␮g/ml). Diagrams represent flow cytometric analysis of CD11c⫹ cells.
Results are representative of three experiments. c, BMDCs derived from
3⫻-␬B-luc transgenic mice were incubated for 2 h with medium alone,
FliC, CpG, or LPS. Luciferase activity was assayed in cell extracts and
normalized to activity in mock-treated BMDC. The results shown are representative of two experiments.
6926
FLAGELLIN/MyD88-SPECIFIC STIMULATION OF Th2-BIASED RESPONSES
CpG, and FliC at various concentrations before monitoring luciferase activity (Fig. 2c) (20). The kinetics of NF-␬B activation were
similar for all TLRs tested, with maximal activity 2 h after addition
of stimuli. Flagellin at concentrations as low as 1 ng/ml enhanced
NF-␬B transcription by ⬃10-fold, whereas the increase with CpG
and LPS was significantly higher. These observations establish that
DC activation by flagellin is mediated by MyD88 and results in
NF-␬B induction, as expected for a TLR5-dependent mechanism.
We next investigated the pattern of cytokine secretion in flagellin-activated DCs. The heterodimeric cytokine IL-12 p70 (formed
of IL-12 p40 and IL-12 p35) is produced upon stimulation of DCs
via various TLRs (31). Interestingly, flagellin did not stimulate any
IL-12 p70 secretion in BMDCs in contrast to CpG (Fig. 3a). However, flagellin induced a significant increase in IL-12 p40 production that is fully dependent on MyD88. Only small quantities of the
proinflammatory cytokines IL-6 and TNF-␣ were produced by
DCs upon incubation with flagellin compared with CpG (Fig. 3a).
In addition, CpG, but not flagellin, triggered systemic IL-12 p70
production after i.v. injection, whereas significant levels of IL-12
p40 were measured with flagellin (Fig. 3b). This cytokine pattern
was not modified by flagellin dose or costimulation, as assessed by
treatment with anti-CD40 mAb (Fig. 3b and data not shown). Although flagellin induces MyD88-specific signaling, mouse DC activation both in vitro and in vivo is not associated with secretion of
the Th1-polarizing cytokine IL-12 p70 and a strong proinflammatory response.
Flagellin induces the development of Th2-type responses
The cytokine IL-12 p70 plays an important role in the development
of Th1 responses (31). The TLR5 ligand flagellin has been proposed to elicit either Th1- or Th2-type immunity in mice (16, 17).
Based on the observations that mouse DCs did not produce any
IL-12 p70 upon flagellin stimulation, we investigated whether
flagellin would promote Th2-biased responses. Mice were immunized s.c. with FliC, OVA, and OVA supplemented with FliC or
CpG oligonucleotide as a Th1-promoting TLR9-dependent adjuvant (32). Ten days later the secretion of IL-4, IL-13, and IFN-␥
was measured in OVA-stimulated lymph node cells (Fig. 4a). T
cells from FliC-immunized animals did not induce any cytokine
release when incubated in vitro with OVA. As expected, immunization with CpG resulted in Th1 polarization, characterized by the
secretion of IFN-␥, but not IL-4 or IL-13. The production of IL-4
and IL-13, cytokines specific for Th2 responses, was only detected
in T cells from animals immunized with FliC and OVA (Fig. 4a).
As reported for other Th2-type inducing stimuli (33), significant
FIGURE 3. Flagellin-mediated cytokine secretion in BMDCs and in vivo. a, BMDCs derived
from C57BL/6 or MyD88⫺/⫺ mice were incubated
for 20 h with medium alone, FliC, or CpG (1 and
10 ␮g/ml, respectively). The secretion of IL-12
p40, IL-12 p70, TNF-␣, or IL-6 was measured by
ELISA. Similar results were obtained with BMDC
from 3⫻-␬B-luc transgenic and C3H/HeJ mice. b,
C3H/HeJ mice (n ⫽ 4) were injected i.v. with
PBS, FliC (10 ␮g), and CpG (50 ␮g), supplemented or not with anti-CD40 mAb (50 ␮g) as
indicated. Serum concentrations of IL-12 p40 and
IL-12 p70 were determined 6 h later.
production of IFN-␥ was found in the group treated with FliC and
OVA. The mouse genotype did not influence the T cell response,
as C57BL/6 mice behaved similarly to BALB/c mice (Fig. 5a).
These data demonstrate that flagellin promotes the development of
Th2 cytokine-producing T cells in vivo.
To further characterize the T cell response, CFSE-labeled,
OVA-specific, transgenic CD4⫹ T cells were transferred into naive
animals and then injected i.v. with PBS, FliC, OVA alone, or OVA
in combination with FliC or CpG (Fig. 4b). Three days later the
proliferation of CFSE-positive cells in spleen or lymph nodes was
analyzed by flow cytometry. Neither FliC nor PBS induced proliferation of transferred CD4⫹ T cells, indicating that flagellin has
no direct effect on OVA-specific T cells. A limited number of cells
underwent divisions in animals treated with OVA alone in contrast
to the study performed with OVA323–339 Th peptide (17). In contrast, coadministration of OVA with flagellin resulted in a dramatic
increase in OVA-specific CD4⫹ T cell proliferation similar to that
observed with CpG and OVA. The absolute number of transgenic
T cells was also markedly increased. These experiments indicate
that flagellin potentates the in vivo activation of Ag-specific CD4⫹ T
cells. OVA-specific CD4⫹ T cells were further isolated from spleens
of immunized animals to analyze IL-4 and IFN-␥ gene expression by
real-time RT-PCR (Fig. 4c). Injection of OVA with CpG resulted in
Th1 polarization, i.e., IFN-␥ expression, but no IL-4 expression. Although IFN-␥ was similarly expressed in all groups, IL-4 expression
was dramatically enhanced in transgenic CD4⫹ T cells isolated from
animals injected with FliC and OVA. Together our observations indicate that flagellin promotes the production of IL-4 by CD4⫹ T cells,
the signature cytokine of a Th2-type response.
Flagellin adjuvant effect requires MyD88
The importance of TLR signaling in Th1 induction is mostly based
on the observation that MyD88-deficient mice are unable to mount
Th1 responses when CFA, which contains TLR agonists, is used as
adjuvant (7). Experiments were conducted to define the role of
MyD88 in the adjuvant properties of flagellin (Fig. 5). MyD88deficient animals and their wild-type counterparts, C57BL/6, were
immunized as described above. As observed with BALB/c mice,
flagellin triggered OVA-specific Th2-biased responses, whereas
CpG elicited Th1 responses (Fig. 5a). In MyD88 animals, flagellin,
like CpG, did not trigger any response in knockout animals (Fig.
5b). In contrast, CFA stimulated the secretion of IL-4 and IL-13,
thereby inducing Th2-type responses (Fig. 5b). We then investigated the Ab response associated with T cell priming in immunized animals. Indeed, when flagellin was coinjected with OVA,
The Journal of Immunology
6927
FIGURE 4. Flagellin promotes the development of Th2-type immune responses in vivo. a, BALB/c mice (n ⫽ 3) were immunized s.c. with FliC (10
␮g), OVA (10 ␮g), or CpG (100 ␮g) as indicated. Ten days later cells from lymph nodes were incubated for 72 h with OVA, and the concentrations of
IL-4, IL-13, and IFN-␥ in supernatants were measured in triplicate. Results are representative of three experiments and are similar to those for spleen. b,
BALB/c mice were adoptively transferred with CFSE-labeled, OVA-specific CD4⫹ T cells. On day 1 mice (n ⫽ 2) were injected with PBS, FliC (10 ␮g),
OVA (100 ␮g), or CpG (100 ␮g). Three days later the CFSE fluorescence of cells from peripheral lymph nodes was analyzed by flow cytometry. Absolute
numbers of transgenic OVA-specific CD4⫹ T cells (⫻106) are indicated on the histograms. The results shown are representative of three experiments and
are similar to those obtained with splenocytes. c, The mRNA levels of IL-4, and IFN-␥ genes were quantified by real-time RT-PCR in OVA-specific CD4⫹
T cells sorted on day 3 from animals injected as described in b. IL-4 and IFN-␥ levels were normalized with respect to the porphobilinogen deaminase gene
and are presented as arbitrary expression units. n.d., not detected. The results shown are representative of two experiments.
the OVA-specific serum IgG response increased ⬎20-fold (Fig.
4c). This effect was effective at doses of flagellin ranging from 100
ng to 30 ␮g/animal, but was totally abolished when flagellin was
digested by trypsin before injection (Figs. 5 and 6). In addition,
flagellin caused a marked increase in anti-OVA IgG1 titers without
significantly increasing the anti-OVA IgG2a response in various
mouse genetic backgrounds, such as C57BL/6, BALB/c, C3H/HeJ,
and NMRI (Fig. 6). In MyD88 animals, flagellin did not potentate any
OVA-specific responses compared with CFA (Fig. 5d). As reported
for innate immunity (14), the effect of flagellin on adaptive immunity,
especially on the development of Th2-biased responses, was therefore
strictly dependent on the adapter molecule MyD88.
Discussion
DCs are instrumental for the development of Th cell responses (2,
3). In particular, it has been postulated that MyD88-dependent
TLR activators promote Th1 responses (7, 8). In this study we
provide evidence that flagellin induces Th2-biased responses in a
6928
FLAGELLIN/MyD88-SPECIFIC STIMULATION OF Th2-BIASED RESPONSES
FIGURE 5. MyD88-dependent stimulation of immunity by flagellin.
C57BL/6 (a and c) or MyD88⫺/⫺ (b and d) mice (n ⫽ 3–5) were immunized s.c. on days 0 and 21 with OVA (10 ␮g) in combination with PBS,
FliC (10 ␮g), trypsin-treated FliC (equivalent to 10 ␮g), CpG (50 ␮g), or
CFA as indicated. a and b, Seven days after boosting, cells from lymph
nodes cells were incubated for 72 h with medium (䡺) or 500 ␮g/ml OVA
(f), and the concentrations of IL-4, IL-13, and IFN-␥ in supernatants were
measured. Similar results were found with splenocytes. c and d, Total
OVA-specific IgG in serum were analyzed by ELISA on day 28. Results
are expressed as the reciprocal geometric mean of Ab titers.
MyD88-dependent manner. We observed that flagellin induces
MyD88-dependent maturation of DCs both in vivo and in vitro. In
vivo, flagellin induces the development of IL-4-producing CD4⫹ T
cells, a hallmark of Th2 polarization. This effect is associated with
the lack of Th1-polarizing cytokine IL-12 p70 production by
flagellin-stimulated DCs.
A major concern in analysis of PRR stimulatory activity is contamination of the PAMP preparation by unrelated ligands, especially for in vivo studies or studies with DCs. Our flagellin preparations were first characterized for the presence of bioactive
flagellin using intestinal epithelial cells as reported previously (19,
28). Although gangliosides have been reported to contribute to
signaling, the flagellin-mediated effect is essentially dependent on
TLR5 stimulation and MyD88 (14, 29, 34, 35). To rule out any
direct contribution of ligands for TLR2, TLR3, TLR4, TLR6, and
TLR9 in the flagellin preparation, we used endotoxin-depleted purified flagellin combined with B cell activation assays, which can
detect these elicitors (data not shown). In addition, flagellin was
fully effective at inducing maturation of splenic DC in mice with
impaired TLR2 and TLR4 signaling. Together these data strongly
support the idea that the bacterial product responsible for biological activity on DCs and adaptive immunity is flagellin, the TLR5
ligand. Although we determined that flagellin promotes Th2 responses in a MyD88-dependent manner. TLR5-deficient mice
should be assayed to define the unique and specific role of TLR5
in this process.
The most striking finding is that Th2-biased differentiation is
associated with MyD88-dependent activation of DCs. Although
CpG and flagellin both induce the production of IFN-␥ by OVAspecific T cells, only flagellin promotes the secretion of Th2-specific cytokines, IL-4 and IL-13. How can signaling in DC via distinct TLRs give rise to different CD4⫹ T cell polarizations?
Several mechanisms, including cell subsets, state of maturation,
and cytokine release, could account for such a difference. CD8␣⫺
DCs have been proposed to be involved in Th2 priming (36). Our
experiments suggest that flagellin-programmed Th2-type differentiation is not controlled by a specific DC subset, as the whole DC
population is activated in vivo (Figs. 1c and 2a). Th2 development
has been associated with strong costimulatory signals provided by
activated DCs (3). After injection of flagellin, splenic DCs were
activated, with the level of costimulatory molecules higher than
that after CpG injection (data not shown). CpG and flagellin have
similar capacities to prime OVA-specific T cells, but have distinct
effects on T cell polarization (Fig. 5). The role of costimulatory
molecules in flagellin-mediated, Th2-type responses is therefore
not conclusive and will require further investigations. Th1/Th2 development can be influenced by the Ag dose provided to DCs. In
the study by McSorley et al. (17), flagellin administration with the
Th peptide OVA323–339 also resulted in IFN-␥ production by CD4
T cells, but IL-4 production was not reported. This Th1-type promoting effect of flagellin, which diverges from our data, suggests
that either the dose of Ag or the processing and presentation of
whole OVA provide essential information for T cell polarization.
The doses of whole OVA used in our work are 220-fold lower than
the dose of peptide used by McSorley et al. (17), assuming that
each OVA molecule will generate a Th peptide. We also observed
a lower activation of DO11.10 cells with OVA compared with that
obtained with OVA323–339 peptide, which alone triggers a strong
proliferation without any microbial stimuli. This discrepancy in
Th1/Th2 balance could therefore be attributed to an Ag dose effect
in which a high Ag dose favors Th1 responses. Alternatively,
flagellin activity on Ag processing and presentation in DCs such as
occurs during immunization with a complete Ag may change the
context in which T cell polarization occurs. Finally, the commitment to a Th1/Th2 phenotype depends on a specific pattern of
cytokines released by DCs. The TLR stimuli tested to date (i.e.,
LPS, peptidoglycan, dsRNA, and CpG) promote and/or enhance
the development of Th1-type responses as well as IL-12 p70 secretion by DCs (31, 32, 37, 38). The lack of IL-12 p70 secretion
and the low levels of IL-6 and TNF-␣ appear to be unique to
flagellin signaling and probably to TLR5 (Fig. 3a). Interestingly,
flagellin-mediated activation of human DCs is also characterized
by a distinct cytokine pattern compared with LPS-mediated activation, including lack of IL-12 p70 secretion (18). This singularity
was already suggested by the moderate signs of inflammation triggered by flagellin compared with other TLR ligands (26, 39). The
failure of DC to produce IL-12 p70 or other Th1-promoting cytokines, such as the cytokines IL-23 and IL-27, during T cell sensitization could directly impair Th1 development, expansion, or
maintenance. Flagellin-dependent Th2 commitment could also rely
on Th1-suppressive mechanisms, such as secretion of IL-10 or
TGF-␤ or production of Th2-inducing cytokines (31).
The Journal of Immunology
6929
FIGURE 6. Development of IgG1 response does not depend on Ag dose or flagellin dose. a, Dose response of OVA. BALB/c mice (n ⫽ 3) were
immunized s.c. on days 0 and 21 with OVA (10 –100 ␮g) alone or in combination with FliC (30 ␮g) as indicated. Similar results were obtained with
C57BL/6 animals. b, Dose response of flagellin. BALB/c mice (n ⫽ 3) were immunized s.c. on days 0 and 21, and NMRI mice were immunized once with
OVA (10 ␮g), alone or in combination with various doses of FliC as indicated. IgG1 and IgG2a responses specific for OVA (a and b) and flagellin (a) were
analyzed by ELISA on day 28. Cross-reactivity between IgG2c and IgG2a was used to specifically detect IgG2c Abs, the signature isotype for Th2-biased
response in C57BL/6 mouse. Histograms correspond to absorbance values obtained with a 1/1000 serum dilution. Individual animals are represented by
circles, and geometric means are shown by horizontal bars.
This study challenges the observations of Means et al. (18) concerning stimulation of isolated splenic mouse DC. Nevertheless,
we also found that splenic DCs incubated ex vivo with flagellin
were not stimulated. Sorted DCs undergo spontaneous maturation
that masks subsequent flagellin activation. It is known that any
treatment of DCs is sufficient to induce their spontaneous maturation, in particular magnetic beads or FACS sorting, collagenase
treatment, movement of culture dishes, or dislodgment of DC clusters (40, 41). Therefore, DC activation was analyzed directly in
vivo after i.v. injection of flagellin or on BMDCs (Figs. 1 and 2).
In our study splenic DCs as well as BMDCs express TLR5 and are
fully activated by flagellin in a MyD88-dependent process. In addition, B lymphocytes that do not have any TLR5 transcripts are
totally unresponsive to flagellin (data not shown). The recent study
by Agrawal et al. (13) proposes that flagellin-activated human DC
express IL-12-p70, in contrast to Means’s work (18). The cellular
and molecular mechanisms of these differences remain to be elucidated. To shed light on the controversial studies of flagellin activity on Th responses, several cues should be explored, including
the purity of flagellin preparations, the influence of Ag (OVA peptide, self-Ag for mixed lymphocyte reactions, or OVA protein), the
relevance of readouts (in vivo or in vitro), the phenotype of DCs
(in vitro- or in vivo-derived DCs, as well as the level of TLR5
expression), and host specificity (rodent vs human).
Recent reports proposed that TLR2 and TLR4 play a role in Th2
differentiation (9 –12). These studies propose that MyD88-independent mechanisms and the dose of TLR agonists contribute to
Th2 polarization. Thus, the difference in T cell differentiation induced by flagellin compared with TLR stimuli may reside in distinct quantitative and qualitative signals. Our data show that in DC
the factors that constitute the core TLR-specific signaling pathway,
i.e., MyD88, NF-␬B, and IL-12 p40 expression, are conserved in
TLR5 if we assume that flagellin-mediated signaling totally depends on TLR5 activation as described by others (14, 29, 42, 43).
We assessed whether the level of DC activation could explain the
difference between CpG and flagellin. Upon flagellin injection,
DCs are activated, and IL-12 p40 is secreted as for TLR-activating
signals, suggesting a similar intensity of signaling. In addition, the
outcome of activation is independent of flagellin dose, suggesting
that TLR5 per se may be sufficient for instructing Th2-biased differentiation. The unique property of flagellin is probably due to a
qualitative difference in the signaling machinery. For instance,
TLRs have individual specificities, characterized by specific signaling complexes composed of MyD88 and accessory adapter
molecules (such as Toll/IL-1R domain-containing adapter protein
(TIRAP), Toll-interacting protein, and Toll/IL-1R domain-containing
adapter-inducing IFN-␤), which define the specificity of TLR2,
TLR3, and TLR4 signaling (44 – 46). TLR5 does not require TIRAP
(45), but whether other adaptors can modulate signal transduction
through TLR5 is yet not known. The host cell glycolipid asialoganglioside monosialic acid 1 can bind flagellin and induce specific signaling, including mobilization of calcium and activation of
mitogen-activated protein kinase (extracellular signal-regulated kinase 1/2) (34). We show that MyD88-dependent processes are essential for the development of flagellin-promoted Th2-type immunity, but TLR5 cooperation with PRRs such as asialo-GM1 may
modulate the outcome of the response, as described recently for
TLR2 and dectin-1 (47). Flagellin therefore represents a unique
model to dissect TLR-associated signaling machinery specialized
in development of Th2-biased responses.
How pathogens influence Th2 differentiation is poorly characterized and has consequently been classified as a default pathway.
In this study flagellin was unambiguously shown to be the PAMP
responsible for Th2 commitment. Usually, products used to study
Th1/Th2 balance are composed of a mixture of microbial structures. For instance, CFA, Toxoplasma extract, or Schistosoma egg
Ags are probably composed of several PAMPs that individually
signal through PRRs promoting Th1 or Th2 responses. Similarly,
commercial Ags or PAMP preparations that are considered pure
PAMPs undoubtedly contain microbial contaminants, especially
LPS. Under these conditions, a hierarchy between the various PRR
signals may account for the preferential instruction of Th1 or Th2
responses. Such a model is supported by recent observations showing that Th1-promoting TLR-specific signals appear to predominate over MyD88-independent Th2-promoting signals (8).
6930
FLAGELLIN/MyD88-SPECIFIC STIMULATION OF Th2-BIASED RESPONSES
In conclusion, our observations support a broader importance of
TLRs in immune responses against pathogens. The combination of
TLRs stimulated by the various PAMPs displayed by microbes
might result in the fine-tuning of adaptive immune responses.
Acknowledgments
We thank Drs. F. Niedergang, D. Finke, M. Rumbo, R. Voyle, M. Corbett,
and H. Acha-Orbea for helpful discussions, and P. Zaech, I. Surbeck, and
F. Hoegger for excellent technical assistance. We are grateful to
Drs. A. Aguzzi and M. Prinz (Institute of Neuropathology, Zurich, Switzerland) for MyD88⫺/⫺ mice, to Dr. K. Hughes (University of Washington,
Seattle, WA) for Salmonella strains, to Dr. A. Aderem for hTLR5 expression plasmid (Institute for Systems Biology, Seattle, WA), to Prof. J. Louis
and Dr. H. Voigt (World Health Organization, Epalinges, Switzerland) for
the IL-4 bioassay, to Prof. R. Steinman (Yale University, New Haven, CT)
for the C4H3 Ab, to Dr. B. Valsazina (Biochemistry Institute, Epalinges,
Switzerland) for DO11.10 SCID mice, and to Dr. A. Wilson (Ludwig Institute, Epalinges, Switzerland) for Abs.
References
1. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of
immunity. Nature 392:245.
2. Moser, M., and K. M. Murphy. 2000. Dendritic cell regulation of TH1-TH2
development. Nat. Immunol. 1:199.
3. Jankovic, D., Z. Liu, and W. C. Gause. 2001. Th1- and Th2-cell commitment
during infectious disease: asymmetry in divergent pathways. Trends Immunol.
22:450.
4. Janeway, C. A., Jr., and R. Medzhitov. 2002. Innate immune recognition. Annu.
Rev. Immunol. 20:197.
5. Barton, G. M., and R. Medzhitov. 2002. Control of adaptive immune responses
by Toll-like receptors. Curr. Opin. Immunol. 14:380.
6. Kawai, T., O. Adachi, T. Ogawa, K. Takeda, and S. Akira. 1999. Unresponsiveness of MyD88-deficient mice to endotoxin. Immunity 11:115.
7. Schnare, M., G. M. Barton, A. C. Holt, K. Takeda, S. Akira, and R. Medzhitov.
2001. Toll-like receptors control activation of adaptive immune responses. Nat.
Immunol. 2:947.
8. Jankovic, D., M. C. Kullberg, S. Hieny, P. Caspar, C. M. Collazo, and A. Sher.
2002. In the absence of IL-12, CD4⫹ T cell responses to intracellular pathogens
fail to default to a Th2 pattern and are host protective in an IL-10⫺/⫺ setting.
Immunity 16:429.
9. Re, F., and J. L. Strominger. 2001. Toll-like receptor 2 (TLR2) and TLR4 differentially activate human dendritic cells. J. Biol. Chem. 276:37692.
10. Pulendran, B., P. Kumar, C. W. Cutler, M. Mohamadzadeh, T. Van Dyke, and
J. Banchereau. 2001. Lipopolysaccharides from distinct pathogens induce different immune responses in vivo. J. Immunol. 167:5067.
11. Kaisho, T., K. Hoshino, T. Iwabe, O. Takeuchi, T. Yasui, and S. Akira. 2002.
Endotoxin can induce MyD88-deficient dendritic cells to support Th2 cell differentiation. Int. Immunol. 14:695.
12. Eisenbarth, S. C., D. A. Piggott, J. W. Huleatt, I. Visintin, C. A. Herrick, and
K. Bottomly. 2002. Lipopolysaccharide-enhanced, Toll-like receptor 4-dependent
T helper cell type 2 responses to inhaled antigen. J. Exp. Med. 196:1645.
13. Agrawal, S., A. Agrawal, B. Doughty, A. Gerwitz, J. Blenis, T. Van Dyke, and
B. Pulendran. 2003. Cutting edge: different Toll-like receptor agonists instruct
dendritic cells to induce distinct Th responses via differential modulation of extracellular signal-regulated kinase-mitogen-activated protein c-Fos. J. Immunol.
171:4984.
14. Hayashi, F., K. D. Smith, A. Ozinsky, T. R. Hawn, E. C. Yi, D. R. Goodlett,
J. K. Eng, S. Akira, D. M. Underhill, and A. Aderem. 2001. The innate immune
response to bacterial flagellin is mediated by Toll-like receptor 5. Nature
410:1099.
15. Smith, K. D., E. Andersen-Nissen, F. Hayashi, K. Strobe, M. A. Bergman,
S. L. Barrett, B. T. Cookson, and A. Aderem. 2003. Toll-like receptor 5 recognizes a conserved site on flagellin protofilament formation and bacterial motility.
Nat. Immunol. 16:16.
16. Parish, C. R. 1996. Immune deviation: a historical perspective. Immunol. Cell.
Biol. 74:449.
17. McSorley, S. J., B. D. Ehst, Y. Yu, and A. T. Gewirtz. 2002. Bacterial flagellin
is an effective adjuvant for CD4⫹ T cells in vivo. J. Immunol. 169:3914.
18. Means, T. K., F. Hayashi, K. D. Smith, A. Aderem, and A. D. Luster. 2003. The
Toll-like receptor 5 stimulus bacterial flagellin induces maturation and chemokine production in human dendritic cells. J. Immunol. 170:5165.
19. Sierro, F., B. Dubois, A. Coste, D. Kaiserlian, J. P. Kraehenbuhl, and J. C. Sirard.
2001. Flagellin stimulation of intestinal epithelial cells triggers CCL20-mediated
migration of dendritic cells. Proc. Natl. Acad. Sci. USA 98:13722.
20. Carlsen, H., J. O. Moskaug, S. H. Fromm, and R. Blomhoff. 2002. In vivo imaging of NF-␬B activity. J. Immunol. 168:1441.
21. Lutz, M. B., N. Kukutsch, A. L. Ogilvie, S. Rossner, F. Koch, N. Romani, and
G. Schuler. 1999. An advanced culture method for generating large quantities of
highly pure dendritic cells from mouse bone marrow. J. Immunol. Methods
223:77.
22. Anjuere, F., P. Martin, I. Ferrero, M. L. Fraga, G. M. del Hoyo, N. Wright, and
C. Ardavin. 1999. Definition of dendritic cell subpopulations present in the
spleen, Peyer’s patches, lymph nodes, and skin of the mouse. Blood 93:590.
23. Reis e Sousa, C., and R. N. Germain. 1999. Analysis of adjuvant function by
direct visualization of antigen presentation in vivo: endotoxin promotes accumulation of dendritic cells in the T cell areas of lymphoid tissue. J. Immunol.
162:6552.
24. Favre, N., and P. Erb. 1993. Use of the CTL44 cell line, a derivative of CTL/L
cells, to identify and quantify mouse interleukin-4 by bioassay. J. Immunol. Methods 164:213.
25. Sebastiani, G., G. Leveque, L. Lariviere, L. Laroche, E. Skamene, P. Gros, and
D. Malo. 2000. Cloning and characterization of the murine Toll-like receptor 5
(Tlr5) gene: sequence and mRNA expression studies in Salmonella-susceptible
MOLF/Ei mice. Genomics 64:230.
26. Renshaw, M., J. Rockwell, C. Engleman, A. Gewirtz, J. Katz, and S. Sambhara.
2002. Impaired Toll-like receptor expression and function in aging. J. Immunol.
169:4697.
27. Edwards, A. D., S. S. Diebold, E. M. Slack, H. Tomizawa, H. Hemmi, T. Kaisho,
S. Akira, and C. Reis e Sousa. 2003. Toll-like receptor expression in murine DC
subsets: lack of TLR7 expression by CD8␣⫹ DC correlates with unresponsiveness to imidazoquinolines. Eur. J. Immunol. 33:827.
28. Gewirtz, A. T., P. O. Simon, C. K. Schmitt, L. J. Taylor, C. H. Hagedorn, A. D.
O’Brien, A. S. Neish, and J. L. Madara. 2001. Salmonella typhimurium translocates flagellin across intestinal epithelia, inducing a proinflammatory response.
J. Clin. Invest. 107:99.
29. Gewirtz, A. T., T. A. Navas, S. Lyons, P. J. Godowski, and J. L. Madara. 2001.
Bacterial flagellin activates basolaterally expressed TLR5 to induce epithelial
proinflammatory gene expression. J. Immunol. 167:1882.
30. Kaisho, T., O. Takeuchi, T. Kawai, K. Hoshino, and S. Akira. 2001. Endotoxininduced maturation of MyD88-deficient dendritic cells. J. Immunol. 166:5688.
31. Trinchieri, G. 2003. Interleukin-12 and the regulation of innate resistance and
adaptive immunity. Nat. Rev. Immunol. 3:133.
32. Chu, R. S., O. S. Targoni, A. M. Krieg, P. V. Lehmann, and C. V. Harding. 1997.
CpG oligodeoxynucleotides act as adjuvants that switch on T helper 1 (Th1)
immunity. J. Exp. Med. 186:1623.
33. Launois, P., I. Maillard, S. Pingel, K. G. Swihart, I. Xenarios, H. Acha-Orbea,
H. Diggelmann, R. M. Locksley, H. R. MacDonald, and J. A. Louis. 1997. IL-4
rapidly produced by V␤4V␣8 CD4⫹ T cells instructs Th2 development and susceptibility to Leishmania major in BALB/c mice. Immunity 6:541.
34. McNamara, N., A. Khong, D. McKemy, M. Caterina, J. Boyer, D. Julius, and
C. Basbaum. 2001. ATP transduces signals from ASGM1, a glycolipid that functions as a bacterial receptor. Proc. Natl. Acad. Sci. USA 98:9086.
35. Ogushi, K. I., A. Wada, T. Niidome, T. Okuda, R. Llanes, M. Nakayama,
Y. Nishi, H. Kurazono, K. D. Smith, A. Aderem, et al. 2004. Gangliosides act as
co-receptors for Salmonella enteritidis FliC and promote FliC-induction of human ␤-defensin-2 expression in Caco-2 cells. J. Biol. Chem. 279:12213.
36. Maldonado-Lopez, R., T. De Smedt, P. Michel, J. Godfroid, B. Pajak,
C. Heirman, K. Thielemans, O. Leo, J. Urbain, and M. Moser. 1999. CD8␣⫹ and
CD8␣⫺ subclasses of dendritic cells direct the development of distinct T helper
cells in vivo. J. Exp. Med. 189:587.
37. Edwards, A. D., S. P. Manickasingham, R. Sporri, S. S. Diebold, O. Schulz,
A. Sher, T. Kaisho, S. Akira, and E. S. C. Reis. 2002. Microbial recognition via
Toll-like receptor-dependent and -independent pathways determines the cytokine
response of murine dendritic cell subsets to CD40 triggering. J. Immunol.
169:3652.
38. Boonstra, A., C. Asselin-Paturel, M. Gilliet, C. Crain, G. Trinchieri, Y. J. Liu, and
A. O’Garra. 2003. Flexibility of mouse classical and plasmacytoid-derived dendritic cells in directing T helper type 1 and 2 cell development: dependency on
antigen dose and differential Toll-like receptor ligation. J. Exp. Med. 197:101.
39. Liaudet, L., K. G. Murthy, J. G. Mabley, P. Pacher, F. G. Soriano, A. L. Salzman,
and C. Szabo. 2002. Comparison of inflammation, organ damage, and oxidant
stress Salmonella enterica serovar Muenchen flagellin and serovar lipopolysaccharide. Infect. Immun. 70:192.
40. Gatti, E., M. A. Velleca, B. C. Biedermann, W. Ma, J. Unternaehrer,
M. W. Ebersold, R. Medzhitov, J. S. Pober, and I. Mellman. 2000. Large-scale
culture and selective maturation of human Langerhans cells from granulocyte
colony-stimulating factor-mobilized CD34⫹ progenitors. J. Immunol. 164:3600.
41. Pierre, P., S. J. Turley, E. Gatti, M. Hull, J. Meltzer, A. Mirza, K. Inaba,
R. M. Steinman, and I. Mellman. 1997. Developmental regulation of MHC class
II transport in mouse dendritic cells. Nature 388:787.
42. Moors, M. A., L. Li, and S. B. Mizel. 2001. Activation of interleukin-1 receptorassociated kinase by Gram-negative flagellin. Infect. Immun. 69:4424.
43. Weinmann, A. S., D. M. Mitchell, S. Sanjabi, M. N. Bradley, A. Hoffmann,
H. C. Liou, and S. T. Smale. 2001. Nucleosome remodeling at the IL-12 p40
promoter is a TLR-dependent, Rel-independent event. Nat. Immunol. 2:51.
44. Zhang, G., and S. Ghosh. 2002. Negative regulation of Toll-like receptor-mediated signaling by Tollip. J. Biol. Chem. 277:7059.
45. Horng, T., G. M. Barton, R. A. Flavell, and R. Medzhitov. 2002. The adaptor
molecule TIRAP provides signalling specificity for Toll-like receptors. Nature
420:329.
46. Oshiumi, H., M. Matsumoto, K. Funami, T. Akazawa, and T. Seya. 2003. TICAM-1, an adaptor molecule that participates in Toll-like receptor 3-mediated
interferon-␤ induction. Nat. Immunol. 4:161.
47. Gantner, B. N., R. M. Simmons, S. J. Canavera, S. Akira, and D. M. Underhill.
2003. Collaborative induction of inflammatory responses by dectin-1 and Tolllike receptor 2. J. Exp. Med. 197:1107.