| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print June 3, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* The Pulmonary Center, Department of Medicine, Boston University School of Medicine, Massachusetts;
Department of Microbiology and Immunology, University of Maryland School of Medicine, Baltimore; and
Department of Pathology, Uniformed Services University of the Health Sciences, Bethesda, Maryland
Correspondence at current address: Dr. Matthew J. Fenton, Division of Pulmonary and Critical Care Medicine, University of Maryland School of Medicine, 685 W. Baltimore Street, MSTF 800, Baltimore, MD 21201-1192. E-mail: mfenton{at}medicine.umaryland.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(IFN-) were similar in macrophages from all three mouse strains, suggesting that BCG growth in the lungs of TLR2-/- mice was a consequence of defective adaptive immunity. Antigenic stimulation of splenocytes from infected wild-type and TLR4-/- mice induced T cell proliferation in vitro, whereas T cells from TLR2-/- mice failed to proliferate. Unexpectedly, activated CD4+ T cells from both TLR-deficient mouse strains secreted little IFN- in vitro compared with control T cells. A role for TLR4 in the control of bacterial growth and IFN- production in vivo was observed only when mice were infected with higher numbers of BCG. Thus, TLR2 and TLR4 appear to regulate distinct aspects of the host immune response against BCG.
Key Words: cytokines • monocytes/macrophages • bacterial (infections) • inflammation • T lymphocytes
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Despite a high degree of similarity between the intracytoplasmic portions of mammalian TLR proteins, activation of different TLR proteins elicits distinct responses from leukocytes. Specifically, stimulation of macrophages via TLR2 induces responses that are quantitatively and qualitatively different from those induced via TLR4 signaling [5
]. Most dramatically, a TLR2 agonist, LPS from Porphyromonas gingivalis, failed to induce significant IL-6, interferon- (IFN-), IL-12p40, and monocyte chemoattractant protein-5 expression compared with the TLR4 agonist, Escherichia coli LPS. More recently, our laboratory found similar differences in the responses of macrophages to activation by TLR2 and TLR4 agonists. Although E. coli LPS and mycobacterial TLR2 agonists could induce similar levels of IL-1ß and tumor necrosis factor 
(TNF-) secretion from murine macrophages, nitric oxide (NO) production was only observed in LPS-stimulated cells [11
]. Subsequent studies showed that these differences were in part a result of TLR4-specific induction of IFN-ß secretion, which acts in an autocrine/paracrine manner to activate signal transducer and activator of transcription (STAT)1 and STAT1-dependent gene expression [12
]. These differences in TLR-mediated macrophage responses in vitro have profound implications for how different TLR proteins contribute in distinct ways to the innate-immune response. Similarly, differences in the development of innate immunity in vivo are likely to affect the subsequent development of acquired immunity.
Among known agonists for mammalian TLR proteins are several molecules derived from mycobacteria. Purified 19-kDa lipoprotein [13 ] and the glycolipids arabinose-capped lipoarabino-mannan [4 ] and dimannosylated phosphatidylinositol (PIM1–2) [11 ] have been found to activate cells in a TLR2-dependent manner. In addition, TLR4 has been shown to mediate cellular activation by live Mycobacterium tuberculosis bacilli [14 ], by phosphatidylinositol mannosides (PIM4–6) from M. tuberculosis [15 ], and by cell-wall preparations purified from Mycobacterium bovis bacillus Calmette-Guerin (BCG) [16 ]. In contrast, no TLR4-agonist activity has yet been detected in Mycobacterium avium [17 ]. Thus, TLR2 and possibly TLR4 would be expected to participate in host responses against mycobacterial infection. The capacities of TLR2 and TLR4 to participate in the activation of leukocytes by mycobacteria and to mediate distinct responses elicited by these cells suggest that TLR proteins may serve distinct roles in the host-immune responses against mycobacteria. To test this hypothesis, TLR2 and TLR4 knockout mice and normal controls were infected with the nonpathogenic mycobacterium M. bovis BCG. These studies revealed that TLR2-deficient mice could not effectively control BCG growth in vivo and exhibited defective innate- and adaptive-immune responses. In contrast, TLR4-deficient mice could control BCG growth in vivo, although antigen-specific T helper cell type 1 (Th1)-type responses were still significantly reduced. When inoculated with higher numbers of bacilli, these TLR4-deficient mice were unable to control BCG growth effectively in vivo in spite of equivalent production of inducible NO synthase (iNOS) protein. Thus, TLR2 and TLR4 regulate innate- and adaptive-immune responses against BCG infection in mice, although the contributions of these TLR proteins to host defense are qualitatively distinct.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Bacteria
M. bovis BCG [Pasteur strain, American Type Culture Collection (ATCC), Manassas, VA] was propagated in sterile, endotoxin-free liquid Middlebrook 7H9 broth supplemented with glycerol, Tween-80, L-asparagine (1 g/L), and an oxo-acid dehydrogenase complex supplement (Fischer Scientific, Springfield, NJ). BCG was grown at 37°C to mid-log phase and stored frozen at -80°C until use in vivo and in vitro. Colony-forming units (CFU) were determined by serial dilution of BCG cultures on Middlebrook 7H11 agar plates. For in vitro experiments, log-phase BCG cultures were dispersed by brief sonication in an ice-cold water bath followed by vortexing. Aggregated mycobacteria were then allowed to sediment at 1 g. Disaggregated bacilli were collected and quantified by spectrophotometry at OD620 immediately before addition to cultured eukaryotic cells. For in vivo infection studies, frozen stocks of BCG were thawed, sonicated briefly in an ice-cold water bath, and diluted into sterile phosphate-buffered saline (PBS). Mice (6–10 weeks old) were injected intraperitoneally (i.p.) with 1 x 106 to 1 x 108 BCG in a final volume of 0.5 ml PBS. Two weeks after infection, mice were killed, whereupon blood and tissues were collected.
Determination of bacterial numbers in lungs
Lungs of infected mice were aseptically removed and placed in sterile 0.025% sodium dodecyl sulfate (SDS) on ice. Lungs were subsequently homogenized using sterile, loose-fitting Teflon pestle and glass tube homogenizers (Glas-Col, Terre Haute, IN). Lung homogenates were briefly sonicated, diluted in sterile water, and cultured on Middlebrook 7H11 agar plates supplemented with 20 µg/ml cycloheximide. Plates were then incubated at 37°C for 14–21 days, after which visible BCG colonies were counted.
In vitro splenocyte restimulation assays
Spleens were aseptically removed from uninfected mice or mice 14 days after BCG infection and were then placed in ice-cold RPMI-1640 medium containing penicillin, streptomycin, HEPES (Gibco, Grand Island, NY), 10% fetal bovine serum (FBS; HyClone, Logan, UT), and 20 µM ß-mercaptoethanol (ß-ME). Splenocyte suspensions were generated using sterile-frosted glass slides and were subsequently treated with ice-cold red blood cell (RBC) lysis buffer (Sigma Chemical Co., St. Louis, MO). Splenocytes (5x105 cells/well) were plated immediately in 96-well U-bottom plates (Fischer Scientific) in 200 µl RPMI 1640 (with 10% FBS, penicillin, streptomycin, HEPES buffer, ß-ME) in the presence or absence of heat-killed (65°C, 30 min) BCG (1x103 bacilli/well). In some experiments, splenocytes from uninfected mice were cultured in wells precoated with an anti-T cell receptor-ß (TCR-ß) antibody (BD PharMingen, San Jose, CA). For proliferation assays, splenocytes were stimulated for 48 h and pulsed during the final 8 h of culture with 1 µCi/well 3H-thymidine (Perkin Elmer Life Sciences, Boston, MA). Cells were then harvested onto filters using an automated cell harvester (Molecular Devices, Sunnyvale, CA). Incorporated radioactivity was quantitated by scintillation fluorography. For cytokine release assays, splenocytes were stimulated for 72 h, and cytokine levels in fresh-culture supernatants were immediately measured using enzyme-linked immunosorbent assay (ELISA) kits (BD Pharmingen), as recommended by the manufacturer.
Selective depletion of CD4+ and CD8+ splenocytes was performed using antibody-coated magnetic beads. Briefly, erythrocyte-depleted splenocytes were pelleted and resuspended in ice-cold PBS. Splenocytes were then incubated with magnetic beads coated with rabbit anti-mouse CD4 or anti-mouse CD8 antibodies (Dynal Biotech, Oslo, Norway) at 4°C for 20 min with gentle agitation. Unbound cells were magnetically separated from bound cells and were washed, and resulting unbound cell numbers were quantified using a hemacytometer. Undepleted splenocytes were manipulated in an identical manner, except that antibody-coated magnetic beads were excluded (i.e., mock-depleted).
Isolation of peritoneal macrophages and infection with live M. bovis BCG
To isolate peritoneal macrophages, mice were injected i.p. with 3 ml LPS-free, sterile thioglycollate (Remel, Lenexa, KS). After 4 days, peritoneal exudate cells were harvested by lavage with sterile RPMI-1640 cell-culture medium. Adherent macrophages were cultured for 3 days before use. Macrophages (5x105 macrophages/well) were infected with live BCG (five bacilli/macrophage) for 24 h in RPMI 1640 containing 10% FBS and cefotaxime (50 µg/ml; Sigma Chemical Co.). Culture supernatants were then collected and filtered through 0.2 µm Spin-X filters (Costar, Corning, NY) to remove any bacteria. Cytokine levels were measured using ELISA kits (BD PharMingen), as recommended by the manufacturer.
In vitro infection and intracellular growth of M. bovis BCG by macrophages
Peritoneal macrophages were prepared as described above and plated (2x105 cells/well) in 48-well tissue-culture plates. Prior to infection, macrophages were washed and cultured in RPMI 1640 containing 10% FBS and 50 µg/ml cefotaxime (Sigma Chemical Co.). Macrophages were incubated with live BCG from log-phase cultures (five bacilli/macrophage) for 4 h and were then washed extensively. Cells were lysed immediately or cultured for an additional 4 days in the absence or presence of 100 ng/ml recombinant murine IFN- (R&D Systems, Minneapolis, MN) before lysis. Culture supernatants were removed, and macrophages were lysed by the addition of 0.025% SDS to each well. Wells were washed twice with lysis buffer, and washes were pooled with the lysates, after which the pooled lysates and washes were briefly sonicated. Lysates were diluted in lysis buffer and then plated on Middlebrook 7H11 agar plates for determination of CFU.
Transfection of HEK 293 cells
HEK 293 cells (ATCC) were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% FBS, penicillin/streptomycin, and L-glutamine. Cells were transfected with the nuclear factor (NF)-
B-dependent endothelial leukocyte adhesion molecule-luciferase (ELAM-luc) reporter plasmid (ref. [20
] from D. Golenbock) and one or more expression plasmids encoding the following genes: ß-galactosidase (ref. [21
] from D. Hwang), murine CD14 (ref. [22
] from D. Golenbock), murine MD-2 (ref. [23
] from D. Golenbock), hemagglutinin (HA)-tagged murine TLR2 (ref. [24
] from A. Aderem), and HA-tagged murine TLR4 (ref. [24
] from A. Aderem). HEK 293 cells were transfected using FuGENE-6 (Roche, Indianapolis, IN), according to the manufacturer’s instructions. Twenty-four hours after transfection, HEK 293 cells were stimulated with live BCG (five bacilli/cell) for 6 h. Cells were then lysed, and ß-galactosidase and luciferase activities were measured using a commercial kit (Promega, Madison, WI), according to the manufacturer’s protocols. E. coli LPS (Sigma Chemical Co.; repurified as described in ref. [5
]) and a synthetic bacterial lipopeptide {S-[2,3-Bis-(palmitoyloxy)-(2-RS)-propyl]-N-palmitoyl-(R)-Cys-(S)-Ser-Lys4-OH, trihydrochloride (Pam3CSK4), EMC Microcollections, Tubingen, Germany} were used as control TLR4 and TLR2 agonists, respectively.
Tissue histology and immunohistochemistry
Sera were collected from saline-injected or BCG-infected C3H/OuJ and TLR4 mutant C3H/HeJ mice before sacrifice 14 days after infection with 1 x 107 bacilli/mouse. ELISA measured levels of circulating TNF- (R&D Systems), IFN-, and IL-12p70 (BD PharMingen), according to the manufacturers’ instructions. Livers from the mice were removed, sliced into 4 mm-thick strips, and fixed in freshly prepared 4% paraformaldehyde. Following fixation, tissues were embedded in paraffin, and 5 µm sections of fixed liver tissues were prepared. Liver sections were deparaffinized and hydrated before hematoxylin and eosin (H&E) staining [25
]. Briefly, sections were stained for 15 min in Mayer’s hematoxylin, washed, and counterstained with eosin <2 min before slides were dehydrated and mounted. Serial sections were also stained for acid-fast bacilli using the Ziehl-Neelsen (ZN) staining method [25
]. Briefly, sections were stained in a carbol fuchsin solution for 30 min, washed, and decolorized in acid alcohol. Subsequently, sections were counterstained with methylene blue (1.4%), washed, and dehydrated before mounting.
Fourteen days after BCG infection, livers from C3H/OuJ and C3H/HeJ mice were removed, fixed in 10%-buffered formalin (Sigma Chemical Co.), and embedded in paraffin, and liver sections (10 mm) were mounted on glass slides. After deparaffinization, endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 30 min at room temperature. For antigen unmasking, the slides were submerged in citrate buffer (pH 6) and heated for 10 min in a microwave oven. The sections were cooled in PBS and subsequently incubated with 3% skim milk for 30 min at room temperature to block nonspecific binding. The rabbit polyclonal antibody against mouse iNOS (Santa Cruz Biotechnology, Santa Cruz, CA) was applied at a 1:750 dilution, and the slides were incubated overnight at 4°C. The next day, the sections were incubated with biotinylated goat anti-rabbit immunoglobulin G (1:200 dilution; Vector Laboratories, Burlingame, CA) and with avidin-biotin-peroxidase complexes (1:100 dilution, Vector Laboratories) for 30 min at room temperature. Peroxidase activity was visualized with 3'-diaminobenzidine tetrahydrochloride (Vector Laboratories), and nuclear counterstaining was performed with Harris hematoxylin. Lastly, the slides were dehydrated and permanently mounted with Permount (Fisher Scientific).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
to control this intracellular growth were subsequently measured. As shown in Figure 1B
, there were no significant differences in the uptake and growth of intracellular BCG in macrophages from the three strains of mice or in the ability of IFN- to suppress intracellular BCG growth in vitro. Therefore, the increased bacterial load observed in BCG-infected TLR2-/- mice relative to controls and TLR4-/- mice does not appear to result from a selective growth advantage of BCG in the macrophages of TLR2-/- mice or from an inherent defect in the bacteriostatic activities of these cells.
TLR2 but not TLR4 mediates a proinflammatory response of macrophages infected with BCG in vitro
As the growth and IFN--dependent suppression of intracellular BCG growth in infected wild-type and TLR-deficient macrophages in vitro were similar, the growth of BCG in the lungs of infected TLR2-/- mice may be secondary to a deficit in inflammatory responses induced in TLR2-/- macrophages by BCG. The relative contribution of TLR2 and TLR4 to activation of macrophage proinflammatory cytokine production following infection with BCG in vitro was further investigated. As shown in Figure 2
, stimulation of peritoneal macrophages from wild-type mice with viable BCG resulted in the secretion of high levels of TNF-, IL-1ß, and IL-6. In contrast, stimulation of macrophages from TLR2-/- mice with BCG resulted in minimal production of these cytokines. These studies revealed that the majority of cytokine-inducing activity in intact, viable BCG is transmitted via TLR2.
|
B-dependent reporter gene (ELAM-luc) and expression plasmids encoding mouse TLR2 or TLR4. In the latter case, cells were also transfected with expression plasmids encoding the TLR4 coreceptors CD14 and MD-2. Twenty-four hours after transfection, cells were stimulated with live BCG (five bacilli/HEK 293 cell), and luciferase activity was measured 6 h later. As shown in Figure 3
, HEK 293 cells expressing neither TLR2 nor TLR4 were nonresponsive to BCG as well as the control TLR2 agonist Pam3CSK4 and the control TLR4 agonist E. coli LPS. However, BCG induced significant reporter gene activity in transfected HEK 293 cells, which express TLR2. In contrast, live BCG did not stimulate HEK 293 cells expressing mouse CD14, MD-2, and TLR4 (which respond robustly to purified E. coli LPS). Thus, BCG activation of cells critically depends on TLR2 expression, although TLR4 is not involved in acute activation of the cells by live BCG. These data suggest that BCG lacks a TLR4 agonist or that such an agonist is not expressed on the surface of viable bacilli.
|
-secreting Th1-type CD4+ T cells (reviewed in ref. [28
]). The development of Th1 immunity also depends on the production of cytokines by macrophages and professional antigen-presenting cells. The finding that TLR2-/- macrophages failed to secrete proinflammatory cytokines in response to BCG challenge in vitro suggested that the development of protective adaptive-immune responses might also be impaired in BCG-infected TLR2-/- mice. This possibility was tested directly using an antigen-dependent splenocyte restimulation assay. Mice were infected with viable BCG (1x106 bacilli/mouse), and spleens were removed 14 days later. Splenocytes were restimulated with heat-killed BCG in vitro, and cytokine production was measured 3 days later. As shown in Figure 4A
, splenocytes from uninfected C57BL/6 wild-type mice did not secrete IFN- in response to stimulation heat-killed BCG in vitro. However, the capacity of antigen-stimulated splenocytes to secrete IFN- correlated with the duration of BCG infection in vivo. Furthermore, IFN- secretion by BCG-restimulated splenocytes was abolished when they were depleted of CD4+ cells but not CD8+ cells (Fig. 4B)
. This suggests that CD4+ T lymphocytes are the predominant IFN--secreting cell type in the splenocyte population of mice 14 days after i.p. infection with BCG. Thus, 2 weeks postinfection represents a useful time-point to evaluate the initiation of normal immune responses in wild-type C57BL/6 mice challenged with BCG.
|
in response to restimulation with heat-killed BCG in vitro, whereas splenocytes from infected TLR2-/- mice secreted little IFN- (Fig. 5A
). However, splenocytes from infected TLR4-/- mice also secreted significantly less IFN- than wild-type splenocytes. This small but consistent level of IFN- production by splenocytes from TLR4-/- mice was abrogated when the splenocyte population was depleted of CD4+ cells before antigenic stimulation (data not shown). There was no inherent defect in the capacities of the TLR-deficient T cells to be stimulated and secrete IFN-, as shown by the ability of a cross-linking anti-TCR-ß antibody to induce similar amounts of IFN- secretion by splenocytes prepared from all three strains of mice (Fig. 5A
, inset). For all genotypes, IL-4 and IL-5 production was undetectable in the splenocyte culture supernatants under all conditions tested (data not shown). Consistent with the results of the IFN- assays, splenocytes from BCG-infected wild-type mice proliferated in response to heat-killed BCG, resulting in 
2.5-fold increase in 3H-thymidine incorporation compared with BCG-stimulated splenocytes from uninfected mice (Fig. 5B)
. Splenocytes from infected TLR4-/- mice also proliferated in response to heat-killed bacteria, indicating that TLR4-/- mice are also capable of generating BCG-specific lymphocytes in vivo. In contrast, splenocytes from infected TLR2-/- mice consistently failed to proliferate in an antigen-dependent manner, suggesting that TLR2-/- mice may not be capable of developing normal acquired immune responses to BCG infection. Together, these data suggest that TLR2- and TLR4-deficient mice exhibit abnormal, adaptive-immune responses following BCG challenge.
|
production (compared with control mice in Fig. 5A
). As IFN- has been shown to be important for control of M. tuberculosis infection in vivo [29
, 30
], the consequences of poor IFN- production in mice lacking functional TLR4 was investigated in greater depth. One possibility was that defective host responses against BCG would become apparent when TLR4 mutant mice were infected with higher numbers of BCG.
To test this possibility, TLR4-/- and wild-type mice were infected i.p. with two different doses of BCG (1x106 and 1x107 bacilli/mouse). Serum levels of IFN- were then compared 14 days later. As shown in Figure 6
, IFN- serum levels were not elevated in control or TLR4-deficient mice infected with the lower dose of BCG. However, when a tenfold-higher inoculum was administered, IFN- levels were significantly increased in the sera of wild-type mice. In contrast, no increase in IFN- levels was observed in the sera of TLR4-/- mice. Thus, a role for TLR4 in host immunity becomes apparent only when mice were infected with higher numbers of BCG.
|
, IFN-, and IL-12p70 were significantly increased in sera of TLR4 normal C3H/OuJ mice following infection. Although similar amounts of TNF- released into the serum of C3H/OuJ and C3H/HeJ mice, lower levels of serum IL-12p70 were associated with the TLR4 mutant C3H/HeJ strain. Consistent with this observation, levels of serum IFN- in infected C3H/HeJ mice were approximately tenfold lower than in sera from BCG-infected C3H/OuJ mice. Thus, TLR4 contributes to the development of strong Th1 responses in TLR4-/- and TLR4 mutant C3H/HeJ mice, but this is only detected when the infecting dose is high.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
to suppress intracellular bacterial growth were similar in wild-type and TLR-deficient macrophages. These findings suggested that the inability of TLR2-/- mice to control BCG growth in the lung was a result, at least in part, of defective adaptive immunity. This possibility was tested using an in vitro antigen restimulation assay to measure secondary T cell activation and cytokine production. These studies revealed that development of antigen-specific CD4+ T cells in vivo is impaired in the infected TLR2-/- mice. In contrast, antigen-specific T cells were generated in BCG-infected TLR4-/- mice, but these cells secreted little IFN- following antigen challenge compared with wild-type T cells. Taken together, these findings support the notion that TLR2 and TLR4 serve distinct roles in the development of host immunity against BCG. The absolute requirement for TLR2 and conversely, the apparent negligible role for TLR4 in mediating macrophage activation by live BCG in vitro (Fig. 2) were unexpected, as one previous study reported that cell-wall preparations from BCG contained a TLR4 agonist activity [16 ]. One possible explanation is that BCG contains a vast excess of TLR2 agonists, relative to TLR4 agonists. Alternatively, these TLR4 agonists might only be present within the bacterial cytosol, or masked within the cell walls of live BCG. This latter possibility is consistent with the inability of live BCG to activate transfected HEK 293 cells in a TLR4-dependent manner (Fig. 3) . However, a role for TLR4 in host responses was revealed in vivo when higher, infective doses of BCG were used (1x107 bacilli/mouse, Figs. 6 and 7 ). This strongly suggests that the presence of higher numbers of BCG in vivo may lead to the expression or release of bacterial TLR4 agonists within infected macrophages at levels sufficient to affect host responses. Alternatively, an endogenous murine TLR4 agonist may affect host responses against BCG, but its contribution to the control of BCG growth in vivo is only apparent when high infective doses BCG are used.
Antigen restimulation assays revealed that CD4+ lymphocytes from infected TLR2-/- mice failed to proliferate and secrete IFN- following antigen challenge in vitro (Figs. 5
and 6)
. These data suggest that BCG did not stimulate detectable expansion of antigen-specific, naïve T cells in vivo. This possible defect in T cell priming in vivo was not a result of an intrinsic defect in TLR2-deficient T cell function, as shown by the finding that these cells could be induced to secrete normal amounts of IFN- following TCR cross-linking (Fig. 6
, inset). Thus, the absence of T cell expansion following infection might arise as a consequence of defective antigen presentation and/or costimulation. This possibility is consistent with our finding that BCG challenge (108 CFU/mouse) strongly increased CD40 and B7.2 surface expression on CD11c+ dendritic cells 6 h after infection in wild-type mice, whereas no increase in the expression of these activation markers was observed in TLR2-/- mice 6 h after BCG infection (data not shown).
Additional antigen restimulation assays revealed that lymphocytes from infected TLR4-/- mice proliferated normally but also exhibited a diminished capacity to secrete high levels of IFN- following antigen challenge in vitro (Figs. 5
and 6)
. These data suggest that BCG could stimulate the expansion of antigen-specific, naïve T cells in vivo but that these T cells did not exhibit a strong Th1 phenotype. BCG infection, like infection with other mycobacteria, induces a strong Th1 host response in humans and mice (reviewed in ref. [28
]). Failure to elicit this Th1 response correlates with progressive infection and chronic disease. In our antigen restimulation assays, wild-type splenocytes produced high levels of IFN- but no detectable IL-4 or IL-5 (data not shown). Moreover, defective IFN- secretion by TLR4-deficient T cells was not a result of an intrinsic defect in the T cells or in the pathways leading to IFN- gene expression, as shown by the capacity of these cells to secrete IFN- following in vitro cross-linking of the TCR-ß chain (Fig. 6
, inset). Collectively, these data suggest that TLR4-deficient T cells do not become strongly polarized to a Th1 phenotype following BCG infection. It is interesting that Kaisho et al. [32
] recently reported that IFN- secretion by T cells was impaired in TLR4-/- mice immunized with keyhole limpit hemocyanin in complete Freund’s adjuvant, a mycobacteria-derived adjuvant, suggesting that the development of a strong Th1 response to myocobacterial challenge may require TLR4 signaling.
Suprisingly, impaired secretion of IFN- by TLR4-deficient T cells in the in vitro antigen-restimulation assay did not correlate with the diminished capacity of the TLR4-/- mice to clear BCG from their lungs (Fig. 1A)
and other organs (data not shown) by 14 days after infection with 1 x 106 bacilli. These findings suggest that sufficient levels of IFN- may still be produced by a variety of different cell types in vivo to control bacterial growth via adaptive-immune responses. In addition, innate-immune responses may play a significant role in controlling BCG growth at relatively low bacterial loads. Nevertheless, our findings raise the possibility that TLR4 could affect the development of adaptive immunity under specific conditions, such as infection with larger numbers of BCG. Using 1 x 107 bacilli per mouse, we found that serum levels of IFN- were not elevated in TLR4-/- mice (compared with infected, wild-type mice, Fig. 6
), findings that were confirmed using normal C3H/OuJ and TLR4 mutant C3H/HeJ mice (Table 1)
. Together, these findings argue for a role for TLR4 in the host response to BCG at higher bacterial loads.
Three previous studies have addressed the role of TLR4 in host responses against M. tuberculosis infection in mice. Chackerian et al. [33 ] reported that C3H/HeJ and C3H/OuJ mice showed no differences in susceptibility to lethal infection with a virulent strain of M. tuberculosis. In these studies, 1 x 106 M. tuberculosis bacilli were given via intravenous injection. This conclusion was supported by the recent findings of Reiling et al. [34 ], who used a high-dose (2x103 CFU) aerosol challenge with M. tuberculosis to show that TLR2- but not TLR4-defective mice were more susceptible to chronic infection, compared with control mice. In contrast, using an aerosol route of infection by 100 bacilli, Abel et al. [15 ] reported that TLR4 mutant C3H/HeJ mice exhibited impaired elimination of the mycobacteria and succumbed more quickly to the infection compared with normal C3H/HeN mice. Furthermore, pulmonary macrophage recruitment and proinflammatory responses against M. tuberculosis were impaired in the C3H/HeJ mice, presumably resulting in chronic infection. An explanation for these contrasting results remains to be determined. In our studies using a progressive model of BCG infection, we also observed that C3H/HeJ mice were unable to eliminate the bacilli effectively compared with wild-type mice. Apparently, the contribution of TLR4 to host responses in vivo can differ, depending on the route of infection and size of inoculum. Nevertheless, the data reported by Abel et al. [15 ] and our data presented here argue that TLR4 can play a protective role in the defense against mycobacterial infection.
We also showed that C3H/HeJ mice infected with a high dose of BCG i.p. were markedly impaired in their ability to control the growth of BCG in the livers, compared with infected C3H/OuJ mice (Fig. 7A) . This increased susceptibility correlated with increased numbers of infiltrating, inflammatory cells, as judged by histochemical staining. It is interesting that iNOS expression in the livers of both infected mouse strains was found to be similar (Fig. 7B) . Thus, differential iNOS expression does not correlate with the susceptibility of C3H/HeJ mice to BCG infection, compared with C3H/OuJ controls. Our results are similar to a previous study showing that iNOS expression in the lungs of C3H/HeN and C3H/HeJ mice infected via aerosol administration with M. tuberculosis was comparable, although the C3H/HeJ mice displayed a greater susceptibility to infection [15 ]. Moreover, our previous finding that M. tuberculosis bacilli could induce NO production by C3H/HeJ macrophages in vitro argues against a role for TLR4 in the induction of iNOS expression [26 ]. Collectively, these data support the possibility that induction of iNOS production in response to mycobacterial challenge in vivo is TLR4-independent.
In summary, our results provide the first direct comparison of the roles of TLR2 and TLR4 in the immune responses elicited by BCG challenge. These studies suggest that TLR2 and TLR4 serve distinct functions in the innate- and adaptive-immune responses against BCG. TLR2 appears to be necessary for the expansion of effector T cells and for the induction of IFN- secretion by these cells. In contrast, TLR4 may be necessary for the development of a normal Th1 response against BCG, particularly when larger bacterial numbers are encountered by the host. These studies may have implications for the importance of TLR proteins in the efficacy of BCG-based vaccines.
|
ACKNOWLEDGEMENTS |
|---|
Received January 17, 2003; revised March 31, 2003; accepted April 11, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
/ß-dependent gene expression in macrophages Nat. Immunol. 3,392-398[CrossRef][Medline]
B binding sites in the human E-selectin gene required for maximal tumor necrosis factor -induced expression Mol. Cell. Biol. 14,5820-5831This article has been cited by other articles:
|
|
 |
|
  T. Yokoi, R. Amakawa, T. Tanijiri, H. Sugimoto, Y. Torii, H. Amuro, Y. Son, K. Tajima, Y.-J. Liu, T. Ito, et al. Mycobacterium bovis Bacillus Calmette-Guerin suppresses inflammatory Th2 responses by inducing functional alteration of TSLP-activated dendritic cells Int. Immunol., October 1, 2008; 20(10): 1321 - 1329. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Marcenaro, B. Ferranti, M. Falco, L. Moretta, and A. Moretta Human NK cells directly recognize Mycobacterium bovis via TLR2 and acquire the ability to kill monocyte-derived DC Int. Immunol., September 1, 2008; 20(9): 1155 - 1167. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. J. Ryu, E. J. Kim, S-H. Lee, S. Y. Kim, G. Y. Suh, M. P. Chung, H. Kim, O. J. Kwon, and W-J. Koh Impaired expression of Toll-like receptor 2 in nontuberculous mycobacterial lung disease Eur. Respir. J., October 1, 2007; 30(4): 736 - 742. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Ferwerda, B. J. Kullberg, D. J. de Jong, S. E. Girardin, D. M. L. Langenberg, R. van Crevel, T. H. M. Ottenhoff, J. W. M. Van der Meer, and M. G. Netea Mycobacterium paratuberculosis is recognized by Toll-like receptors and NOD2 J. Leukoc. Biol., October 1, 2007; 82(4): 1011 - 1018. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Fremond, D. Togbe, E. Doz, S. Rose, V. Vasseur, I. Maillet, M. Jacobs, B. Ryffel, and V. F. J. Quesniaux IL-1 Receptor-Mediated Signal Is an Essential Component of MyD88-Dependent Innate Response to Mycobacterium tuberculosis Infection J. Immunol., July 15, 2007; 179(2): 1178 - 1189. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Dietrich, R. Billeskov, T. M. Doherty, and P. Andersen Synergistic Effect of Bacillus Calmette Guerin and a Tuberculosis Subunit Vaccine in Cationic Liposomes: Increased Immunogenicity and Protection J. Immunol., March 15, 2007; 178(6): 3721 - 3730. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kurtz, K. P. McKinnon, M. S. Runge, J. P.-Y. Ting, and M. Braunstein The SecA2 Secretion Factor of Mycobacterium tuberculosis Promotes Growth in Macrophages and Inhibits the Host Immune Response Infect. Immun., December 1, 2006; 74(12): 6855 - 6864. [Abstract] |
