Immune Responses Elicit Distinct Legionella pneumophila Dendritic ...

2 downloads 0 Views 286KB Size Report
response to live L. pneumophila that was consistent with their cell cycle progression. ... boosting the adaptive immune responses against Legionella infec- tion.
The Journal of Immunology

Dendritic Cells Pulsed with Live and Dead Legionella pneumophila Elicit Distinct Immune Responses1 Toshiaki Kikuchi,2 Takao Kobayashi, Kazunori Gomi, Takuji Suzuki, Yutaka Tokue, Akira Watanabe, and Toshihiro Nukiwa Legionella pneumophila is the causative pathogen of Legionnaires’ disease, which is characterized by severe pneumonia. In regard to the pathophysiology of Legionella infection, the role of inflammatory phagocytes such as macrophages has been well documented, but the involvement of dendritic cells (DCs) has not been clarified. In this study, we have investigated the immune responses that DCs generate in vitro and in vivo after contact with L. pneumophila. Heat- and formalin-killed L. pneumophila, but not live L. pneumophila, induced immature DCs to undergo similar phenotypic maturation, but the secreted proinflammatory cytokines showed different patterns. The mechanisms of the DC maturation by heat- or formalin-killed L. pneumophila depended, at least in part, on Toll-like receptor 4 signaling or on Legionella LPS, respectively. After transfer to naive mice, DCs pulsed with dead Legionella produced serum Ig isotype responses specific for Legionella, leading to protective immunity against an otherwise lethal respiratory challenge with L. pneumophila. The in vivo immune responses required the Ag presentation of DCs, especially that on MHC class II molecules, and the immunity yielded cross-protection between clinical and environmental strains of L. pneumophila. Although the DC maturation was impaired by live Legionella, macrophages were activated by live as well as dead L. pneumophila, as evidenced by the up-regulation of MHC class II. Finally, DCs, but not macrophages, exhibited a proliferative response to live L. pneumophila that was consistent with their cell cycle progression. These findings provide a better understanding of the role of DCs in adaptive immunity to Legionella infection. The Journal of Immunology, 2004, 172: 1727–1734.

L

egionnaires’ disease, caused by Legionella pneumophila, an intracellular Gram-negative bacillus, is clinically manifest as pneumonia (1). The first line of host defense against L. pneumophila infection is provided by mononuclear phagocytes such as monocytes and macrophages, which readily phagocytose the bacteria (1– 4). Humoral immunity plays a secondary role in the host defense, and the presence of a specific opsonizing Ab improves not only the phagocytosis, but also the killing of the organism by phagocytes (1– 4). The involvement of cellular immunity, especially mediated by CD4⫹ Th cells, is also indicated by the clinical observation that Legionnaires’ disease is more common and more severe for individuals with depressed cell-mediated immunity, including transplant recipients, patients given corticosteroid therapy, and patients suffering from acquired immunodeficiency syndrome (1– 4). Although research efforts to clarify the pathophysiology of Legionella infection have mainly shed light on the interactions between macrophages and L. pneumophila, the detailed mechanisms underlying the establishment of adaptive immunity to Legionella remain elusive. Department of Respiratory Oncology and Molecular Medicine, Institute of Development, Aging and Cancer, Tohoku University, Sendai, Japan Received for publication June 16, 2003. Accepted for publication November 18, 2003. 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 These studies were supported, in part, by Public Trust Haraguchi Memorial Cancer Research Fund, Tokyo, Japan; the Mochida Memorial Foundation for Medical and Pharmaceutical Research, Tokyo, Japan; the Smoking Research Foundation, Tokyo, Japan; the Mitsubishi Pharma Research Foundation, Osaka, Japan; and Ministry of Education, Culture, Sports, Science, and Technology, Tokyo, Japan (15012205, 15019010, 15025209, and 15659194). 2 Address correspondence and reprint requests to Dr. Toshiaki Kikuchi, Department of Respiratory Oncology and Molecular Medicine, Institute of Development, Aging and Cancer, Tohoku University, 4-1 Seiryomachi, Aobaku, Sendai 980-8575, Japan. E-mail address: [email protected]

Copyright © 2004 by The American Association of Immunologists, Inc.

Dendritic cells (DCs)3 are APCs with an exquisite ability to interact with T cells and modulate their responses for the generation of immunological memory (i.e., Ag-specific adaptive immunity) (5, 6). In contrast to mononuclear phagocytes, the primary functions of which are aimed at the clearance of invading microorganisms by their avid ingestion, which is known as Ag-nonspecific innate immunity, DCs phagocytose pathogens and subsequently migrate to the lymphoid organs, where DCs display Ags processed and loaded on MHC molecules (5–10). During migration, DCs exhibit drastic changes in their features, termed maturation, including the up-regulation of MHC, costimulatory (e.g., B7-1 and B7-2), adhesion (e.g., ICAM-1), and signaling molecules (e.g., CD40) and the induced production of proinflammatory cytokines, resulting in the improved ability of DCs to activate Agspecific T cells, especially CD4⫹ Th cells, in a MHC class IIrestricted manner (5–10). Thereafter, activated Ag-specific CD4⫹ T cells stimulate the immune effectors (e.g., Ag-specific B cells) (5–10). Accordingly, we hypothesized that DCs also play a key role in boosting the adaptive immune responses against Legionella infection. To evaluate this hypothesis, we pulsed mouse bone marrowderived DCs with live or dead Legionella and assessed their phenotype changes in vitro and their ability to induce anti-Legionella immunity in vivo. The data demonstrate that DCs pulsed with dead L. pneumophila, but not DCs pulsed with live L. pneumophila, undergo phenotypic maturation and, when adoptively transferred, render mice resistant to a lethal pulmonary infection of Legionella. It was also shown that live L. pneumophila impairs the maturation of DCs despite its ability to induce the activation of macrophages.

3

Abbreviations used in this paper: DC, dendritic cell; TLR, Toll-like receptor. 0022-1767/04/$02.00

1728

Materials and Methods Mice Female C57BL/6 (H-2b) and BALB/c mice (H-2d) (Charles River Breeding Laboratories, Atsugi, Japan); A/J (H-2a), C3H/HeN (H-2k), and C3H/HeJ (H-2k) mice (SLC, Hamamatsu, Japan); and MHC class I-deficient (B6.129-B2mtm1) and MHC class II-deficient (B6.129-Abbtm1) mice that had been backcrossed to the C57BL/6 background (Taconic Farms, Germantown, NY), 6 – 8 wk old, were housed under pathogen-free conditions until infection.

Bacterial strains The L. pneumophila strains used in this study included a clinical isolate (referred to in this work as Suzuki strain, serogroup 1) and an environmental isolate (serogroup 5). All Legionella experiments in this study used the Suzuki strain of L. pneumophila, unless otherwise noted. The Escherichia coli strain 25922 was obtained from American Type Culture Collection (Manassas, VA). L. pneumophila was grown on buffered charcoal yeast extract agar plates or buffer yeast extract broth, as previously published (11). Luria-Bertani broth was used as the growth medium for E. coli. The bacteria were washed three times and suspended in sterile PBS, pH 7, before use, and the concentration was adjusted spectrophotometrically.

DC and macrophage preparations DCs were generated from mouse bone marrow precursors in complete RPMI 1640 medium (10% heat-inactivated FBS, 2 mM L-glutamine, 100 ␮g/ml streptomycin, and 100 U/ml penicillin) with 10 ng/ml mouse rGMCSF (R&D Systems, Minneapolis, MN) and 2 ng/ml mouse rIL-4 (R&D Systems), as described previously (12–14). Mouse peritoneal macrophages were obtained by washing the peritoneal cavity with PBS and eliminating nonadherent cells in the following incubation of the peritoneal exudate cells, as described previously (15). DCs and macrophages used in this study were prepared from A/J mice, unless otherwise noted.

DCs AND L. pneumophila spheres (Beckman Coulter, Miami, FL) were used to calibrate the count of the cells. Cells were analyzed on an EPICS XL cytometer with EXPO32 ADC software (Beckman Coulter). Dead cells and debris were excluded from the analysis by gating on the appropriate forward scatter, side scatter, and propidium iodide-staining profile. To determine the percentage of stained cells above the isotype control staining, 1% of false positive events was accepted in the control Ab. For cell cycle analysis, DCs were fixed with 70% ethanol, and their DNA was stained with 20 ␮g/ml propidium iodide. Stained cells were analyzed on the flow cytometer with MultiCycle software (Phoenix Flow Systems, San Diego, CA). The concentrations of specific cytokines released into the medium were measured using ELISA kits for mouse IL-12p40, TNF-␣, IL-6, or IL-1␤ (BioSource International, Camarillo, CA).

Statistical analysis All data are reported as mean ⫾ SE, unless otherwise noted. Statistical comparison was made using the two-tailed Student’s t test, and a value of p ⬍ 0.05 was accepted as indicating significance. Survival evaluation was conducted using Kaplan-Meier analysis.

Results Surface phenotype of L. pneumophila-pulsed DCs Heat- and formalin-killed L. pneumophila induced DC maturation in vitro, but live L. pneumophila did not (Fig. 1). Compared with naive DCs, the percentages of positive cells for surface markers characteristic of mature DCs, including I-Ak (MHC class II; Fig. 1A), CD40 (Fig. 1B), CD54 (ICAM-1; Fig. 1C), CD80 (B7-1; Fig.

Bacterial pulsing, immunization, and infection of mice DCs were incubated with heat-killed (80°C, 10 min), formalin-killed (2%, 15 min), or live L. pneumophila or heat-killed E. coli at a ratio of 50 bacteria per 1 DC for 3 h at 37°C. The amount of bacteria used for pulsing DCs was determined by microscopic observation that 89.5% of DCs took up fluorescence-labeled L. pneumophila, when pulsed at this ratio. DCs were incubated for a further 1 h with 50 ␮g/ml gentamicin sulfate to kill the remaining bacteria, washed extensively with PBS, and then cultured in complete RPMI 1640 medium at 37°C for the in vitro studies. Sterilization of the pulsed DCs was confirmed by monitoring the growth of extracellular and intracellular bacteria during the 1-wk culture period. Where indicated, DCs were incubated with bacteria or 100 ng/ml LPS from E. coli serotype 055:B5 (Sigma-Aldrich, St. Louis, MO) in the presence or absence of 10 ␮g/ml polymyxin B (Sigma-Aldrich). In some experiments, macrophages were pulsed with L. pneumophila, as described for DCs. For immunization, the pulsed DCs were injected i.v. at 5 ⫻ 105 cells per mouse. Three weeks after the immunization, lethal respiratory infection with L. pneumophila was induced, as described previously (12, 16). Briefly, anesthetized mice were placed in a supine position, and 50 ␮l of 5 ⫻ 107 CFU L. pneumophila was inoculated via the trachea into the lung. All animals were monitored daily for 14 days after the inoculation. Obviously moribund mice were sacrificed, and this was recorded as the time of death.

Anti-Legionella Abs Abs against L. pneumophila in sera were assessed by ELISA in microtiter plates (Nalge Nunc International, Rochester, NY) coated with 107 CFU gentamicin-killed (50 ␮g/ml, 1 h) L. pneumophila per well. Serum was serially diluted in TBS (pH 7.4) containing 0.5% BSA for the analysis. Rabbit IgG Abs against mouse IgM, IgG1, IgG2a, IgG2b, IgG3, or IgA were used as secondary Abs for the isotype determination of antiLegionella Abs, and tertiary alkaline phosphatase-conjugated Ab against rabbit IgG was used for the detection (all Abs were from Pierce Biotechnology, Rockford, IL). After the addition of p-nitrophenyl phosphate substrate solution (Pierce Biotechnology), absorbance of the reaction was measured at 405 nm.

Flow cytometric analysis and cytokine ELISA For DC surface Ag expression, DCs were incubated with FITC- or PEconjugated mAbs against I-Ak (MHC class II, clone 11-5.2), CD40 (clone 3/23), CD54 (ICAM-1, clone 3E2), CD80 (B7-1, clone 16-10A1), and CD86 (B7-2, clone GL1), or appropriate isotype-matched control Abs (BD PharMingen, San Jose, CA). For DC proliferation, Flow-Count Fluoro-

FIGURE 1. Flow cytometric analysis of L. pneumophila-pulsed DCs. DCs from A/J mice were pulsed with heat-killed, formalin-killed, or live L. pneumophila at a ratio of 50 bacteria per 1 DC, and analyzed 2 days later for accessory molecule expression by flow cytometry (bold line): A, I-Ak (MHC class II); B, CD40; C, CD54 (ICAM-1); D, CD80 (B7-1); E, CD86 (B7-2). Overlay histogram (gray, filled) in each panel depicts naive DCs as a control. The percentages of stained cells above isotype control staining are shown in each panel.

The Journal of Immunology

1729

1D), and CD86 (B7-2; Fig. 1E), were much higher in DCs pulsed with heat-killed L. pneumophila (I-Ak⫹, 48.0 vs 28.5%; CD40⫹, 56.8 vs 48.0%; CD54⫹, 36.6 vs 23.8%; CD80⫹, 12.2 vs 2.7%; CD86⫹, 24.9 vs 11.4%). Similar results were achieved in DCs pulsed with formalin-killed L. pneumophila (I-Ak⫹, 45.2 vs 28.5%; CD40⫹, 52.4 vs 48.0%; CD54⫹, 29.8 vs 23.8%; CD80⫹, 8.0 vs 2.7%; CD86⫹, 18.9 vs 11.4%). In contrast, the percentage of expression of these surface markers was strikingly lower in DCs pulsed with live L. pneumophila than that in naive DCs (I-Ak⫹, 12.5 vs 28.5%; CD40⫹, 12.8 vs 48.0%; CD54⫹, 1.7 vs 23.8%; CD80⫹, 1.0 vs 2.7%; CD86⫹, 1.6 vs 11.4%). The degree of phenotypic changes induced by L. pneumophila was correlated with the amount of Legionella used for pulsing DCs (Table I). The results were not influenced by purification of MACS-sorted CD11c⫹ DCs from the DC culture, and similar results were observed with DCs prepared from C57BL/6 mice (data not shown). Cytokines secreted from L. pneumophila-pulsed DCs ELISA analyses demonstrated that heat- and formalin-killed L. pneumophila induced DCs to secrete proinflammatory cytokines with distinct (IL-12 and TNF-␣) or similar (IL-6 and IL-1␤) intensities, but, with the exception of IL-1␤ secretion, live L. pneumophila did not (Fig. 2). Pulsing of DCs with heat- or formalinkilled L. pneumophila resulted in IL-12 induction peaking at 36 h (687 ⫾ 12 pg/ml) or 24 h (301 ⫾ 29 pg/ml), respectively (Fig. 2A). Heat- or formalin-killed L. pneumophila also stimulated TNF-␣ secretion from DCs, and the TNF-␣ levels were 235 ⫾ 25 or 514 ⫾ 35 pg/ml, respectively, at the peak (heat-killed Legionella, 48 h; formalin-killed Legionella, 12 h; Fig. 2B). In response to heat- and formalin-killed L. pneumophila, IL-6 secretion from DCs reached a similar peak (2964 ⫾ 153 and 2750 ⫾ 259 pg/ml, respectively) at the time point of 24 h and then stayed at about the same level for another 24 h (Fig. 2C). Heat- and formalin-killed L. pneumophila caused similarly enhanced secretion of IL-1␤ from DCs, the highest being 265 ⫾ 22 and 233 ⫾ 19 pg/ml, respectively, at the time point of 2 h (Fig. 2D). IL-1␤ secretion from DCs was enhanced also by live L. pneumophila, and the highest level was 121 ⫾ 16 pg/ml at the time point of 6 h (Fig. 2D). Apart from IL-1␤, DCs pulsed with live L. pneumophila secreted relatively low levels of cytokines, including IL-12, TNF-␣, and IL-6, during the culture period as did naive DCs (live Legionella, IL-12 ⬍ 91 pg/ml, TNF-␣ ⬍ 65 pg/ml, IL-6 ⬍ 609 pg/ml; naive DCs, IL-12 ⬍ Table I. Phenotype changes of DCs pulsed with L. pneumophila at different ratiosa Positive Cells (%) k

Ratio (Legionella:DC)

I-A (MHC class II)

CD54 (ICAM-1)

CD86 (B7-2)

Naive DCs Heat-killed L. pneumophila 5:1 50:1 100:1 Formalin-killed L. pneumophila 5:1 50:1 100:1 Live L. pneumophila 5:1 50:1 100:1

31.7

22.2

18.2

43.6 48.4 49.0

34.1 40.7 39.8

30.6 37.1 41.1

39.8 46.8 48.9

30.3 37.3 40.9

23.3 28.2 34.1

21.7 0.7 0.5

13.9 2.2 1.1

9.6 2.6 1.9

a The study was similar to that in Fig. 1, but DCs were pulsed with L. pneumophila at several ratios of bacteria to DC.

FIGURE 2. Cytokines released from L. pneumophila-pulsed DCs. DCs from A/J mice were pulsed with heat-killed (f), formalin-killed (E), or live L. pneumophila (⌬). Controls included naive DCs pulsed without L. pneumophila (䡺). The concentrations of cytokines secreted into the medium at varying times were assayed by ELISA. Data represent means ⫾ SE (n ⫽ 3 per data point). A, IL-12; B, TNF-␣; C, IL-6; D, IL-1␤.

41 pg/ml, TNF-␣ ⬍ 83 pg/ml, IL-6 ⬍ 358 pg/ml, IL-1␤ ⬍ 58 pg/ml; Fig. 2). LPS dependency and independency To assess whether the DC maturation induced by heat- and formalin-killed L. pneumophila depends on LPS, an immunostimulatory component of L. pneumophila, we supplemented cultures for bacterial pulsing with polymyxin B, a well-characterized pharmacologic LPS antagonist (17) (Fig. 3, A and B). Compared with DCs pulsed in the absence of polymyxin B, the maturation of DCs pulsed with formalin-killed L. pneumophila was suppressed by the presence of polymyxin B in the DC priming, whereas the maturation of DCs pulsed with heat-killed L. pneumophila was not adversely affected, as indicated by the percentage of CD86⫹ cells on flow cytometric analyses (heat-killed Legionella, 33.3 vs 32.2%; formalin-killed Legionella, 30.1 vs 37.0%; Fig. 3A). In control experiments using E. coli LPS, polymyxin B clearly suppressed the LPS-stimulated DC maturation, and the proportion of CD86⫹ cells markedly decreased (17.2 vs 43.5%; Fig. 3A). Similar results were observed with expression of other surface markers examined, except for CD40 (Table II). No apparent changes were observed in the CD86 expression of naive and live Legionella-pulsed DCs, regardless of polymyxin B supplementation (data not shown). Consistent with this, IL-12 secretion from DCs in response to formalin-killed L. pneumophila and E. coli LPS was inhibited by the addition of polymyxin B compared with each control (i.e., pulsing without polymyxin B), whereas IL-12 secretion from all other DCs was not affected by polymyxin B supplementation (formalin-killed Legionella, p ⬍ 0.005; LPS, p ⬍ 0.05; naive DCs, p ⬎ 0.7; heatkilled Legionella, p ⬎ 0.9; live Legionella, p ⬎ 0.6; Fig. 3B). These data suggest that the DC maturation by formalin-killed L. pneumophila depends, at least in part, on Legionella LPS, but the DC maturation by heat-killed L. pneumophila does not at all.

1730

DCs AND L. pneumophila Table II. Surface phenotype of DCs pulsed with L. pneumophila in the presence or absence of polymyxin Ba Positive Cells (%)

Heat-killed L. pneumophila PB ⫺ PB ⫹ Formalin-killed L. pneumophila PB ⫺ PB ⫹ E. coli LPS PB ⫺ PB ⫹

I-Ak

CD40

CD54

CD80

47.5 51.6

61.2 60.3

43.7 52.3

40.9 41.3

51.5 44.3

56.5 56.3

40.0 37.2

37.6 30.8

55.4 32.3

77.7 54.1

60.1 41.8

51.5 26.6

a The study was similar to that in Fig. 3A, but Legionella-pulsed DCs were analyzed for I-Ak (MHC class II), CD40, CD54, and CD80 expression.

FIGURE 3. Effects of LPS and TLR4 signaling on L. pneumophilainduced DC maturation. A and B, DCs were pulsed with heat-killed, formalin-killed, live L. pneumophila, or LPS in the presence of either medium alone (⫺; A, gray filled histogram; B, 䡺) or the LPS antagonist polymyxin B (⫹; A, bold line; B, f), and analyzed 2 days later for CD86 expression by flow cytometry (A) or IL-12 release into the medium by ELISA (B). C and D, DCs prepared from C3H/HeN mice (wild type; C, gray filled histogram; D, 䡺) or C3H/HeJ mice (TLR4 mutant; C, bold line; D, f) were pulsed with heat-killed, formalin-killed, live L. pneumophila, or LPS, and analyzed 2 days later for CD86 expression by flow cytometry (C) or IL-12 release into the medium by ELISA (D). For A and C, the percentages of stained cells above the isotype control staining are shown in each panel. For B and D, controls included naive DCs, and data represent means ⫾ SE (n ⫽ 3 per data point).

Toll-like receptor 4 (TLR4) signaling dependency and independency We next investigated whether the DC response to heat- and formalin-killed L. pneumophila depends on TLR4 signaling, which has been implicated in the response to E. coli LPS, by using TLR4 mutant mice (C3H/HeJ, unresponsive to E. coli LPS because of a point mutation of the TLR4 gene affecting the Toll/IL-1R domain) and wild-type mice (C3H/HeN) (18 –21) (Fig. 3, C and D). When DCs prepared from TLR4 mutant and wild-type mice were pulsed with E. coli LPS, CD86 expression on TLR4 mutant DCs was severely impaired, resulting in a decreased percentage of CD86⫹ cells compared with that on wild-type DCs (5.6 vs 43.4%; Fig. 3C). Similar results were achieved in TLR4 mutant or wild-type DCs pulsed with heat-killed L. pneumophila, but the impairment of CD86 expression on TLR4 mutant DCs was less impressive (34.0 vs 37.9%; Fig. 3C). In contrast, when pulsed with formalin-killed L. pneumophila, DCs from TLR4 mutant mice displayed slightly more CD86⫹ cells than those from wild-type mice (38.8 vs 36.1%; Fig. 3C). Similar results were observed with CD54 expression (TLR4 mutant vs wild type: heat-killed Legionella, 46.5 vs 53.4%; formalin-killed Legionella, 59.3 vs 51.7%; E. coli LPS, 45.7 vs

57.3%; data not shown). No apparent differences between TLR4 mutant and wild-type DCs were observed in the CD86 expression of naive and live Legionella-pulsed DCs (data not shown). This was relevant to the level of IL-12 secretion from TLR4 mutant and wild-type DCs (Fig. 3D). Enhanced secretion of IL-12 from wildtype DCs pulsed with heat-killed L. pneumophila and E. coli LPS was significantly abrogated in TLR4 mutant DCs (heat-killed Legionella, p ⬍ 0.05; LPS, p ⬍ 0.005), whereas the levels of IL-12 secretion from wild-type and TLR4 mutant DCs were comparable in all other groups (naive DCs, p ⬎ 0.3; formalin-killed Legionella, p ⬎ 0.7; live Legionella, p ⬎ 0.05; Fig. 3D). These data indicate that at least one pathway mediated by the TLR4 signaling is involved in the signaling through which heatkilled L. pneumophila acts on DCs, but that the TLR4 signaling is not required for the response of DCs to formalin-killed L. pneumophila at all. Taken together with our data showing that DC maturation caused by heat- and formalin-killed L. pneumophila is LPS independent and dependent, respectively (Fig. 3, A and B), these findings led to the conclusion that the effects of L. pneumophila LPS are not mediated by signaling through TLR4. Anti-Legionella Ab responses of immunized mice The in vivo immune response by DCs pulsed with L. pneumophila was assessed by determining the serum level of Legionella-specific Abs (Fig. 4). A/J mice immunized with DCs pulsed with heatkilled, formalin-killed, and live L. pneumophila produced significant amounts of all serum anti-Legionella Ab isotypes examined compared with mice immunized with naive DCs, except for IgA (OD405 at 1:90: IgM, p ⬍ 0.01; IgG1, p ⬍ 0.05; IgG2a, p ⬍ 0.05; IgG2b, p ⬍ 0.05; IgG3, p ⬍ 0.01; IgA, p ⬎ 0.1; Fig. 4). The Ab levels induced by DCs pulsed with heat- and formalin-killed L. pneumophila were comparable in almost all isotypes we assayed (OD405 at 1:90: IgM, p ⬍ 0.05; IgG1, p ⬎ 0.7; IgG2a, p ⬎ 0.4; IgG2b, p ⬎ 0.8; IgG3, p ⬎ 0.3; IgA, p ⬎ 0.3), and were higher than those generated by DCs pulsed with live L. pneumophila, despite significant differences only in IgM and IgG3 isotypes (OD405 at 1:90: IgM, p ⬍ 0.05; IgG1, p ⬎ 0.1; IgG2a, p ⬎ 0.05; IgG2b, p ⬎ 0.4; IgG3, p ⬍ 0.05; IgA, p ⬎ 0.3; Fig. 4). Similar results were achieved in C57BL/6 mice immunized with Legionella-pulsed DCs (data not shown). In vivo protective effects of L. pneumophila-pulsed DCs The observed difference in anti-Legionella levels was relevant to the protection of immunized mice against a lethal bronchopulmonary infection of L. pneumophila (Fig. 5, A and B). Immunization of C57BL/6 mice with DCs pulsed with heat- or formalin-killed L.

The Journal of Immunology

FIGURE 4. In vivo Legionella-specific Ab responses of mice immunized with DCs pulsed with L. pneumophila. A/J mice were immunized with heat-killed (f), formalin-killed (E), or live L. pneumophila (⌬). Controls included mice immunized with naive DCs (䡺). Two weeks after immunization, each isotype of anti-Legionella Abs was assessed in serum using a standard ELISA protocol. Results represent means of three mice per data point. A, IgM; B, IgG1; C, IgG2a; D, IgG2b; E, IgG3; and F, IgA.

pneumophila provided 100 or 90% survival against a lethal challenge with L. pneumophila, respectively ( p ⬍ 0.0001, compared with naive DCs; Fig. 5A). In contrast, only 30% of mice immunized with DCs pulsed with live L. pneumophila were protected, and immunization with naive DCs provided no survival against the lethal challenge of L. pneumophila ( p ⬍ 0.05, DCs pulsed with live Legionella compared with naive DCs; Fig. 5A). Although similar results were achieved with immunized BALB/c mice (data not shown), modest protection against a subsequent Legionella challenge was observed with immunized A/J mice, which are known to be genetically susceptible to Legionella infection and are useful as a mouse model of human Legionnaires’ disease (22, 23) (Fig. 5B). Immunization of A/J mice with DCs pulsed with heat- or formalinkilled L. pneumophila led to 40 or 30% survival, respectively, whereas no mice undergoing immunization of DCs pulsed with live L. pneumophila or naive DCs survived the infection of L. pneumophila ( p ⬍ 0.0001 or p ⬍ 0.05, DCs pulsed with heat- or formalin-killed Legionella compared with naive DCs, respectively; p ⬎ 0.8, DCs pulsed with live Legionella compared with naive DCs; Fig. 5B). Requirements for DC expression of MHC class I and II molecules To assess the role of MHC Ag presentation by DCs in the induction of protective immune responses in vivo, DCs were prepared from wild-type, MHC class I-deficient, or MHC class II-deficient C57BL/6 mice for pulsing with Legionella and were used to im-

1731

FIGURE 5. Immunization with DCs pulsed with L. pneumophila against lethal L. pneumophila respiratory infection. A and B, C57BL/6 mice (A) or A/J mice (B) were immunized i.v. with syngenic DCs pulsed with heat-killed (f), formalin-killed (E), or live L. pneumophila Suzuki strain (⌬). C, MHC class I and II presentations of DCs. Wild-type C57BL/6 mice were immunized with DCs pulsed with heat-killed L. pneumophila Suzuki strain using DCs prepared from MHC class I-deficient (E), class II-deficient (⌬), or wild-type C57BL/6 mice (f). D, Immunity specific to the pathogen used for pulsing DCs. A/J mice were immunized with syngenic DCs pulsed with heat-killed L. pneumophila Suzuki strain (clinical strain (f), L. pneumophila environmental strain (E), or E. coli (⌬)). For all parts, controls included mice immunized with naive DCs (䡺). All respiratory challenges were with L. pneumophila Suzuki strain 3 wk after the immunization. Survival was recorded as the percentage of surviving animals (n ⫽ 10 mice per group).

munize wild-type C57BL/6 mice 3 wk before intratracheal instillation of L. pneumophila (Fig. 5C). When pulsed with heat-killed L. pneumophila, MHC class II-deficient DCs provided no protection against Legionella challenge, which was not significantly different from the result with naive wild-type DCs ( p ⬎ 0.6; Fig. 5C). Immunization using MHC class I-deficient DCs pulsed with heatkilled L. pneumophila conferred some protection against lethal Legionella infection compared with that using naive wild-type DCs ( p ⬍ 0.05), but no mice survived until the end of the experiment on day 14 ( p ⬍ 0.0001, class I-deficient DCs compared with Legionella-pulsed wild-type DCs; Fig. 5C). Microbe-specific immunity The microbe specificity of protective immunity developed by Legionella-pulsed DCs was ascertained using different strains of L. pneumophila and another Gram-negative bacterium, E. coli (Fig. 5D). In these studies, A/J mice were immunized with DCs pulsed with a heat-killed clinical strain of L. pneumophila (Suzuki strain), heat-killed environmental strain of L. pneumophila, or heat-killed E. coli 3 wk before challenge with the L. pneumophila clinical strain (Suzuki strain). DC immunization with clinical and environmental strains of L. pneumophila provided comparable protection ( p ⬎ 0.5), resulting in 40 and 30% survival against a lethal challenge of Legionella ( p ⬍ 0.01 and p ⬍ 0.05, compared with immunization using naive DCs, respectively; Fig. 5D). In contrast,

1732 the survival of mice immunized with E. coli-pulsed DCs against pulmonary Legionella infection was not improved ( p ⬎ 0.7, compared with naive DCs; Fig. 5D). As a control, when mice were challenged with a lethal intratracheal infection of E. coli, no protective effect was observed with Legionella-pulsed DCs ( p ⬎ 0.8, DCs pulsed with either strain of Legionella compared with naive DCs; data not shown). Surface phenotype of Legionella-pulsed macrophages Given that live L. pneumophila was shown to abrogate DC maturation in contrast to dead L. pneumophila, we next investigated whether a similar pattern of macrophage responses to live and dead L. pneumophila also occurs by flow cytometric analyses for their expression of I-Ak, MHC class II (Fig. 6A). Compared with naive macrophages, the percentage of positive cells for I-Ak increased in

FIGURE 6. Responses of DCs and macrophages to live L. pneumophila. A, Flow cytometric analysis of MHC class II expression of L. pneumophila-pulsed macrophages. Macrophages were pulsed with heat-killed, formalin-killed, or live L. pneumophila, and analyzed 2 days later for I-Ak (MHC class II) expression by flow cytometry (bold line). The overlay histogram (gray, filled) in each panel depicts naive macrophages as a control. The percentages of stained cells above the isotype control staining are shown in each panel. B, Proliferation of L. pneumophila-pulsed DCs. DCs were pulsed with heat-killed (f), formalin-killed (E), or live L. pneumophila (⌬). The number of viable cells was determined by flow cytometry, as described in Materials and Methods. The data are presented as the mean percentage increase or decrease over the baseline of triplicate wells. Controls included naive DCs pulsed without L. pneumophila (䡺). C, Proliferation of L. pneumophila-pulsed macrophages. This study was similar to that in B, but macrophages were used instead of DCs. D, Cell cycle distribution of L. pneumophila-pulsed DCs. DCs were pulsed with heat-killed (f), formalin-killed (3), or live L. pneumophila (u). The DNA content was analyzed 6 h later by propidium iodide staining and flow cytometry. Controls included naive DCs pulsed without L. pneumophila (䡺). Results represent the percentages of cells in the S or G2/M phase (mean ⫾ SE, n ⫽ 3 per data point). For all panels, DCs and macrophages were prepared from A/J mice.

DCs AND L. pneumophila macrophages pulsed with both heat- and formalin-killed L. pneumophila (heat-killed Legionella, 73.9 vs 49.7%; formalinkilled Legionella, 62.2 vs 49.7%; Fig. 6A), like DCs pulsed with heat- and formalin-killed L. pneumophila (Fig. 1A). When pulsed with live L. pneumophila, macrophages up-regulated MHC class II expression, as indicated by the increased proportion of I-Ak⫹ cells (55.2 vs 49.7%; Fig. 6A), unlike DCs pulsed with live L. pneumophila, which strikingly down-regulated the I-Ak expression (Fig. 1A). DC-restricted proliferative responses to live Legionella To further explore the responses of DCs and macrophages to L. pneumophila, we examined and characterized their proliferative phenotype, finding that DCs, but not macrophages, markedly proliferated only when pulsed with live L. pneumophila (Fig. 6, B and C). In this context, DCs generated from A/J mice with GM-CSF and IL-4 were unpulsed or pulsed with heat-killed, formalin-killed, or live L. pneumophila and then cultured for 7 days without GMCSF and IL-4 to analyze their proliferation. Only pulsing DCs with live L. pneumophila induced a significant increase in viable cell counts, and cell yields reached 269 ⫾ 19% of the starting cell numbers at day 6 ( p ⬍ 0.0001, DCs pulsed with live Legionella compared with all other DCs; Fig. 6B). As was seen with naive DCs, the number of DCs pulsed with heat- and formalin-killed L. pneumophila was similarly reduced starting from the initiation of the culture, but the reduction was not greater than that of naive DCs ( p ⬎ 0.2, compared between heat- and formalin-killed Legionella; p ⬍ 0.001, DCs pulsed with heat- and formalin-killed Legionella compared with naive DCs; Fig. 6B). Similar results were achieved using DCs prepared from C57BL/6 mice (data not shown). In contrast, examination of macrophages after the pulse with Legionella indicated decreased numbers of viable macrophages regardless of what type of Legionella they had been pulsed with ( p ⬎ 0.05, naive macrophages compared with all other macrophages; Fig. 6C). The proliferation of DCs in response to live L. pneumophila was associated with significant increases in the proportion of live Legionella-pulsed DCs in the S and G2/M phase fractions ( p ⬍ 0.05 for both S and G2/M, DCs pulsed with live Legionella compared with all other DCs; Fig. 6D). These data suggest that pulsing DCs with live L. pneumophila rescues the growth inhibition and the G0/G1 cell cycle arrest, both of which are induced by the depletion of growth cytokines. The increased proportion of live Legionellapulsed DCs in S and G2/M phases became obscured 6 days after the initiation of the culture (data not shown).

Discussion Although considerable effort has been focused on host immune mechanisms that mediate protection against infection of L. pneumophila, to date there have been no reports demonstrating the involvement of DCs. Thus, in the present study, we hypothesized that DCs have a potential role in the development of adaptive immunity to Legionella infection. Several pieces of evidence described in the present work support this hypothesis. Adoptively transferred DCs pulsed ex vivo with dead L. pneumophila, but not live L. pneumophila, elicited Legionella-specific Ig isotype responses in vivo that might have contributed to the protection against subsequent lethal challenge with L. pneumophila in a microbe-specific manner. This in vivo effect was correlated with DC phenotypic maturation that was promoted through a TLR4-dependent or TLR4-independent signaling pathway by dead L. pneumophila. Studies conducted using knockout mice suggested that the in vivo protective efficacy of Legionella-pulsed DCs was mediated by MHC class II-restricted CD4⫹ Th cell immune responses. Taken

The Journal of Immunology together with these findings, the inability of live L. pneumophila to trigger DC maturation and its macrophage-stimulating ability suggested that macrophages or DCs might be central to innate or adaptive immunity, respectively, against Legionella infection. Legionella elicits immune responses mediated by CD4⫹ Th cells, especially Th1 cells, after infection, resulting in the generation of adaptive immunity to reinfection, as in the case of other intracellular pathogens (e.g., Listeria, Mycobacteria, Toxoplasma, Leishmania, and Chlamydia) (24 –29). In particular, IFN-␥, which is secreted from Th1 cells primed in primary infection, is considered to be critical for host resistance to reinfection, because IFN-␥ activates mononuclear phagocytes such as monocytes and macrophages, which are primary effector cells against L. pneumophila (30 –32). Humoral Ab responses also serve for the adaptive immunity against Legionella reinfection, as suggested by the following evidence: a CD4⫹ T cell-mediated type-specific Ab response, initially IgM followed by IgG, occurs in patients with Legionnaires’ disease, and anti-Legionella Abs promote the killing of L. pneumophila by activated phagocytes (1– 4). Although this understanding of the mechanisms mediating Ag-specific host protection against Legionella highlights the importance of priming CD4⫹ T cells specific for Legionella protein epitopes, which APCs take on the function of CD4⫹ T cell priming has yet to be investigated. Hence, in this study, we evaluated the concept that DCs capture L. pneumophila that has been killed by innate immunity (i.e., mononuclear phagocytes), mature to enhance the T cell stimulatory capacity, and present Legionella-derived Ags to CD4⫹ T cells together with costimulators (e.g., costimulatory molecules and cytokines), thereby enabling CD4⫹ T cells to induce Legionellaspecific adaptive immunity, as described above. Consistent with this concept, the present study demonstrated that DCs pulsed with dead L. pneumophila, but not DCs pulsed with live L. pneumophila, underwent maturation with the up-regulation of MHC class II, costimulatory (B7-1 and B7-2), adhesion (ICAM-1), and signaling molecules (CD40), and the increased production of proinflammatory cytokines (IL-12, TNF-␣, IL-6, and IL-1␤), and that, when adoptively transferred, dead Legionella-pulsed DCs induced Legionella-specific protective immunity in a manner dependent on MHC class II Ag presentation to CD4⫹ Th cells. In regard to the type of Th cell response conferred by Legionella-capturing DCs, Th1 immune responses are most likely generated in vivo, because DCs pulsed with dead L. pneumophila released large amounts of IL-12, which is the most crucial cytokine for the differentiation of naive T cells into IFN-␥-producing Th1 cells in vivo (33). Further evidence comes from the observation that adoptively transferred DCs pulsed with L. pneumophila generated elevated levels of anti-Legionella serum Abs in Th1-linked IgG2a and IgG3 isotypes, which well correlated with the protection of DC-immunized mice against Legionella challenge. However, a similar elevation was also observed in the levels of Th2linked IgG1 serum Abs, and it is therefore not clear whether Legionella-capturing DCs provoke a strong predominance of Th1 immune responses in vivo. In this context, the Th2-mediated immunity induced by Legionella-pulsed DCs may be in part responsible for the results observed in the DC immunization-challenge experiments, because beneficial roles of Th2-related as well as Th1-related immune responses against Legionella infection have been demonstrated (34). Recent studies have shown that pathogen-associated molecular patterns, components commonly found on the pathogen that are not normally found in the mammalian host, are potent activators of APCs, including macrophages and DCs, and that they are recognized by ligand-specific TLRs; for example, E. coli LPS signals through TLR4, whereas the cell wall components of Gram-positive

1733 bacteria and peptidoglycans from Staphylococcus aureus signal through TLR2 (20, 21). Until recently, the TLR engagement of components derived from L. pneumophila was uncertain. In the current study, analyses of DC stimulation using LPS antagonist polymyxin B and TLR4 mutant DCs revealed that TLR4 or Legionella LPS was involved in the DC maturation by heat- or formalin-killed L. pneumophila, respectively, thus suggesting that Legionella LPS is not recognized by TLR4. These findings support recent observations that LPS derived from L. pneumophila required TLR2 rather than TLR4 to stimulate mouse bone marrow granulocytes (35). Although detailed molecular mechanisms of polymyxin B-Legionella LPS interactions await further studies, it is conceivable that the TLR2 signaling may be, in part, responsible for the DC maturation triggered by formalinkilled L. pneumophila. The present study demonstrated that, in response to live L. pneumophila, DCs exhibited impaired maturation in contrast to the activation seen in macrophages, suggesting that the innate immune system, such as macrophages, first attack the bacteria to kill them, and this attack is essential for DCs to establish the adaptive immunity against Legionella infection. Although exposure to pathogens generally gives rise to DC maturation, exceptions to this have been reported (7–10). Some pathogens have been shown to manage to avoid inducing DC maturation. For example, the parasite Leishmania mexicana transforms into the noninfectious amastigote form for persistent infection (36). Other pathogens are known to possess mechanisms for inhibiting DC maturation. This is illustrated by HSV-1 and vaccinia virus; HSV-1 is thought to act intracellularly, targeting a signal transduction pathway related to DC maturation, and vaccinia may secrete proteins that inhibit cytokines involved in DC maturation (37, 38). Similarly, erythrocytes infected with the malaria parasite Plasmodium falciparum have been shown to interfere with DC maturation (39). Although the mechanisms by which the capture of live L. pneumophila allows DCs to promote the cell cycle for their proliferation are elusive, the phenomenon may accompany impaired maturation of these DCs. These findings will fuel further work toward understanding the interaction between the immune system and pathogens.

Acknowledgments We thank Drs. K. Tateda and K. Yamaguchi (Toho University School of Medicine, Tokyo, Japan) for the gift of Legionella strains, and B. Bell for reading the manuscript.

References 1. Yu, V. L. 2000. Legionella pneumophila (Legionnaires’ disease). In Principles and Practice of Infectious Diseases, Vol. 2. G. L. Mandell, J. E. Bennett, and R. Dolin, eds. Churchill Livingstone, Philadelphia, p. 2424. 2. Brieland, J. K., N. C. Engleberg, G. B. Huffnagle, D. G. Remick, and J. C. Fantone. 2000. Host pathogen interactions in Legionnaires’ disease: lessons learned from a murine animal model. Immunopharmacology 48:249. 3. Friedman, H., Y. Yamamoto, and T. W. Klein. 2002. Legionella pneumophila pathogenesis and immunity. Semin. Pediatr. Infect. Dis. 13:273. 4. Friedman, H., Y. Yamamoto, C. Newton, and T. Klein. 1998. Immunologic response and pathophysiology of Legionella infection. Semin. Respir. Infect. 13:100. 5. Banchereau, J., F. Briere, C. Caux, J. Davoust, S. Lebecque, Y. J. Liu, B. Pulendran, and K. Palucka. 2000. Immunobiology of dendritic cells. Annu. Rev. Immunol. 18:767. 6. Mellman, I., and R. M. Steinman. 2001. Dendritic cells: specialized and regulated antigen processing machines. Cell 106:255. 7. Palucka, K., and J. Banchereau. 2002. How dendritic cells and microbes interact to elicit or subvert protective immune responses. Curr. Opin. Immunol. 14:420. 8. Rescigno, M., and P. Borrow. 2001. The host-pathogen interaction: new themes from dendritic cell biology. Cell 106:267. 9. Moll, H., and C. Berberich. 2001. Dendritic cells as vectors for vaccination against infectious diseases. Int. J. Med. Microbiol. 291:323. 10. Reis e Sousa, C., A. Sher, and P. Kaye. 1999. The role of dendritic cells in the induction and regulation of immunity to microbial infection. Curr. Opin. Immunol. 11:392. 11. Tateda, K., T. A. Moore, J. C. Deng, M. W. Newstead, X. Zeng, A. Matsukawa, M. S. Swanson, K. Yamaguchi, and T. J. Standiford. 2001. Early recruitment of

1734

12.

13.

14.

15.

16.

17.

18.

19.

20. 21. 22.

23.

24.

25.

neutrophils determines subsequent T1/T2 host responses in a murine model of Legionella pneumophila pneumonia. J. Immunol. 166:3355. Kikuchi, T., S. Worgall, R. Singh, M. A. S. Moore, and R. G. Crystal. 2000. Dendritic cells genetically modified to express CD40 ligand and pulsed with antigen can initiate antigen-specific humoral immunity independent of CD4⫹ T cells. Nat. Med. 6:1154. Kikuchi, T., M. A. S. Moore, and R. G. Crystal. 2000. Dendritic cells modified to express CD40 ligand elicit therapeutic immunity against preexisting murine tumors. Blood 96:91. Kikuchi, T., M. Maemondo, K. Narumi, K. Matsumoto, T. Nakamura, and T. Nukiwa. 2002. Tumor suppression induced by intratumor administration of adenovirus vector expressing NK4, a 4-kringle antagonist of hepatocyte growth factor, and naive dendritic cells. Blood 100:3950. Nakamura, A., Y. Mori, K. Hagiwara, T. Suzuki, T. Sakakibara, T. Kikuchi, T. Igarashi, M. Ebina, T. Abe, J. Miyazaki, et al. 2003. Increased susceptibility to LPS-induced endotoxin shock in secretory leukoprotease inhibitor (SLPI)deficient mice. J. Exp. Med. 197:669. Kikuchi, T., and R. G. Crystal. 2001. Antigen-pulsed dendritic cells expressing macrophage-derived chemokine elicit Th2 responses and promote specific humoral immunity. J. Clin. Invest. 108:917. Wiese, A., M. Mu¨ nstermann, T. Gutsmann, B. Lindner, K. Kawahara, U. Za¨ hringer, and U. Seydel. 1998. Molecular mechanisms of polymyxin B-membrane interactions: direct correlation between surface charge density and self-promoted transport. J. Membr. Biol. 162:127. Poltorak, A., X. He, I. Smirnova, M. Y. Liu, C. Van Huffel, X. Du, D. Birdwell, E. Alejos, M. Silva, C. Galanos, et al. 1998. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 282:2085. Qureshi, S. T., L. Larivie`re, G. Leveque, S. Clermont, K. J. Moore, P. Gros, and D. Malo. 1999. Endotoxin-tolerant mice have mutations in Toll-like receptor 4 (Tlr4). J. Exp. Med. 189:615. Takeda, K., T. Kaisho, and S. Akira. 2003. Toll-like receptors. Annu. Rev. Immunol. 21:335. Werling, D., and T. W. Jungi. 2003. Toll-like receptors linking innate and adaptive immune response. Vet. Immunol. Immunopathol. 91:1. Brieland, J., P. Freeman, R. Kunkel, C. Chrisp, M. Hurley, J. Fantone, and C. Engleberg. 1994. Replicative Legionella pneumophila lung infection in intratracheally inoculated A/J mice: a murine model of human Legionnaires’ disease. Am. J. Pathol. 145:1537. Diez, E., S. H. Lee, S. Gauthier, Z. Yaraghi, M. Tremblay, S. Vidal, and P. Gros. 2003. Birc1e is the gene within the Lgn1 locus associated with resistance to Legionella pneumophila. Nat. Genet. 33:55. Cooper, A. M., A. D. Roberts, E. R. Rhoades, J. E. Callahan, D. M. Getzy, and I. M. Orme. 1995. The role of interleukin-12 in acquired immunity to Mycobacterium tuberculosis infection. Immunology 84:423. Tripp, C. S., M. K. Gately, J. Hakimi, P. Ling, and E. R. Unanue. 1994. Neutralization of IL-12 decreases resistance to Listeria in SCID and C.B-17 mice: reversal by IFN-␥. J. Immunol. 152:1883.

DCs AND L. pneumophila 26. Gazzinelli, R. T., M. Wysocka, S. Hayashi, E. Y. Denkers, S. Hieny, P. Caspar, G. Trinchieri, and A. Sher. 1994. Parasite-induced IL-12 stimulates early IFN-␥ synthesis and resistance during acute infection with Toxoplasma gondii. J. Immunol. 153:2533. 27. Mattner, F., K. Di Padova, and G. Alber. 1997. Interleukin-12 is indispensable for protective immunity against Leishmania major. Infect. Immun. 65:4378. 28. Perry, L. L., K. Feilzer, and H. D. Caldwell. 1997. Immunity to Chlamydia trachomatis is mediated by T helper 1 cells through IFN-␥-dependent and -independent pathways. J. Immunol. 158:3344. 29. Brieland, J. K., D. G. Remick, M. L. LeGendre, N. C. Engleberg, and J. C. Fantone. 1998. In vivo regulation of replicative Legionella pneumophila lung infection by endogenous interleukin-12. Infect. Immun. 66:65. 30. Gebran, S. J., Y. Yamamoto, C. Newton, T. W. Klein, and H. Friedman. 1994. Inhibition of Legionella pneumophila growth by ␥ interferon in permissive A/J mouse macrophages: role of reactive oxygen species, nitric oxide, tryptophan, and iron(III). Infect. Immun. 62:3197. 31. Salins, S., C. Newton, R. Widen, T. W. Klein, and H. Friedman. 2001. Differential induction of ␥ interferon in Legionella pneumophila-infected macrophages from BALB/c and A/J mice. Infect. Immun. 69:3605. 32. Skerrett, S. J., and T. R. Martin. 1994. Intratracheal interferon-␥ augments pulmonary defenses in experimental legionellosis. Am. J. Respir. Crit. Care Med. 149:50. 33. De Becker, G., V. Moulin, F. Tielemans, F. De Mattia, J. Urbain, O. Leo, and M. Moser. 1998. Regulation of T helper cell differentiation in vivo by soluble and membrane proteins provided by antigen-presenting cells. Eur. J. Immunol. 28:3161. 34. Newton, C., S. McHugh, R. Widen, N. Nakachi, T. Klein, and H. Friedman. 2000. Induction of interleukin-4 (IL-4) by Legionella pneumophila infection in BALB/c mice and regulation of tumor necrosis factor ␣, IL-6, and IL-1␤. Infect. Immun. 68:5234. 35. Girard, R., T. Pedron, S. Uematsu, V. Balloy, M. Chignard, S. Akira, and R. Chaby. 2003. Lipopolysaccharides from Legionella and Rhizobium stimulate mouse bone marrow granulocytes via Toll-like receptor 2. J. Cell Sci. 116:293. 36. Bennett, C. L., A. Misslitz, L. Colledge, T. Aebischer, and C. C. Blackburn. 2001. Silent infection of bone marrow-derived dendritic cells by Leishmania mexicana amastigotes. Eur. J. Immunol. 31:876. 37. Engelmayer, J., M. Larsson, M. Subklewe, A. Chahroudi, W. I. Cox, R. M. Steinman, and N. Bhardwaj. 1999. Vaccinia virus inhibits the maturation of human dendritic cells: a novel mechanism of immune evasion. J. Immunol. 163:6762. 38. Salio, M., M. Cella, M. Suter, and A. Lanzavecchia. 1999. Inhibition of dendritic cell maturation by herpes simplex virus. Eur. J. Immunol. 29:3245. 39. Urban, B. C., D. J. Ferguson, A. Pain, N. Willcox, M. Plebanski, J. M. Austyn, and D. J. Roberts. 1999. Plasmodium falciparum-infected erythrocytes modulate the maturation of dendritic cells. Nature 400:73.