Systemic activation of dendritic cells by Toll-like receptor ... - Nature

© 2006 Nature Publishing Group http://www.nature.com/natureimmunology

ARTICLES

Systemic activation of dendritic cells by Toll-like receptor ligands or malaria infection impairs cross-presentation and antiviral immunity Nicholas S Wilson1,2,3,6,7, Georg M N Behrens1,6,7, Rachel J Lundie1,2,3,7, Christopher M Smith1,2,3, Jason Waithman4, Louise Young1,3, Simon P Forehan1,3, Adele Mount1,3, Raymond J Steptoe5, Ken D Shortman1,2, Tania F de Koning-Ward1, Gabrielle T Belz1,2, Francis R Carbone2,4, Brendan S Crabb1,2, William R Heath1,2 & Jose A Villadangos1,2 The mechanisms responsible for the immunosuppression associated with sepsis or some chronic blood infections remain poorly understood. Here we show that infection with a malaria parasite (Plasmodium berghei) or simple systemic exposure to bacterial or viral Toll-like receptor ligands inhibited cross-priming. Reduced cross-priming was a consequence of downregulation of crosspresentation by activated dendritic cells due to systemic activation that did not otherwise globally inhibit T cell proliferation. Although activated dendritic cells retained their capacity to present viral antigens via the endogenous major histocompatibility complex class I processing pathway, antiviral responses were greatly impaired in mice exposed to Toll-like receptor ligands. This is consistent with a key function for cross-presentation in antiviral immunity and helps explain the immunosuppressive effects of systemic infection. Moreover, inhibition of cross-presentation was overcome by injection of dendritic cells bearing antigen, which provides a new strategy for generating immunity during immunosuppressive blood infections.

Dendritic cells (DCs) have a variety of features that make them highly effective agents for detecting pathogens and inducing immune responses. DCs are very efficient at all forms of endocytosis, which enables them to sample their environment for the presence of pathogens or virus-infected cells1. In addition, DCs process such exogenous antigens intracellularly and present them to CD4 T cells via major histocompatibility complex (MHC) class II molecules1. DCs also show an unusual specialization in their MHC class I presentation pathway. Although most cells use their MHC class I molecules to present peptides derived from endogenously synthesized proteins, DCs have the capacity to deliver exogenous antigens to the MHC class I pathway, a phenomenon known as cross-presentation2. Many populations of DCs have been described in vivo, but the DEC205+CD8+ subset (CD8 DC) are particularly efficient at crosspresenting antigens2. Although cross-presentation underlies the ability to induce cytotoxic T lymphocyte (CTL) immunity (cross-priming) against some tumor and cellular antigens, its function in viral immunity is debated2–6. To detect pathogens, DCs use a variety of receptors, including those of the Toll-like receptor (TLR) family, which recognize pathogenassociated molecular patterns such as bacterial lipopolysaccharide

(LPS), methylated oligonucleotides (CpG) and viral double-stranded RNA (‘mimicked’ with the synthetic analog polyinosinicpolycytidylic acid (poly(I) poly(C)))7. After they detect pathogen cues, DCs undergo a process of ‘maturation’, which is characterized by transient upregulation of phagocytosis followed by its downregulation8–10, and increased expression of MHC and T cell costimulatory molecules; the latter enables mature DCs to activate naive T cells and thus initiate immunity1. The immune-activating effects of TLR ligands have prompted their use in vaccine formulations11. Paradoxically, however, overt infection of the blood, such as with sepsis or malaria, can be immunosuppressive despite the abundant provision of TLR ligands12–15. The basis for such immunosuppression remains poorly understood. Here we show that DCs activated in vivo by exposure to TLR ligands or malaria parasites do not cross-present subsequently encountered antigens or mount CTL immunity to viruses. Although cross-priming was considerably impaired, the mature DCs retained their capacity to present viral antigens via the endogenous MHC class I processing pathway. Furthermore, neither treatment with TLR ligand nor malaria infection impaired the proliferation of T cells that encountered DCs presenting their cognate antigen in vivo, challenging the idea of

1The Walter and Eliza Hall Institute of Medical Research and 2The Cooperative Research Centre for Vaccine Technology, Parkville, Victoria 3050, Australia. 3Department of Medical Biology and 4Department of Microbiology and Immunology, University of Melbourne, Parkville, Victoria 3010, Australia. 5The Centre for Immunology and Cancer Research, The University of Queensland, Woolloongabba, Queensland 4102, Australia. 6Present addresses: CSL, Parkville, Victoria 3052, Australia (N.S.W.), and Division of Clinical Immunology, Hannover Medical School, Hannover 3623, Germany (G.M.N.B.). 7These authors contributed equally to this work. Correspondence should be addressed to B.S.C. ([email protected]), W.R.H. ([email protected]) or J.A.V. ([email protected]).

Received 6 October 2005; accepted 9 December 2005; published online 15 January 2006; doi:10.1038/ni1300

NATURE IMMUNOLOGY

VOLUME 7

NUMBER 2

FEBRUARY 2006

165

ARTICLES

a

100

Dividing cells (% relative to control)

Dividing cells Control

50

25

0

0

CFSE

c gBT-I T cells proliferating (× 103)

200

150

100

50

0 Untr. CpG act-mOVA DC

5

10

20

CpG injected (nmol)

b Dividing cells (% relative to control)


CpG

75

CD8 CD8, CpG Non-CD8 Non-CD8, CpG

140 120 100 80 60 40

Uninfected

20 0 0

B6 DC

2

5

7

10

Total DC (×103)

substantial involvement of regulatory T cells in the mechanism of immunosuppression. Our findings correlate viral immunity with functional cross-presentation, demonstrate a mechanism likely to contribute to the immunosuppressive effects of sepsis or malaria infection, and provide a rational approach for overcoming such suppression by treatments that bypass the processing requirements for cross-presentation. We also emphasize a hazardous side effect of the systemic administration of TLR ligands, which may nonetheless be beneficial for the treatment of immunopathologies associated with viral infection and autoimmunity. Finally, our results indicate that vaccine formulations should be designed so that antigens and adjuvants are administered in close physical and temporal association to prevent premature inactivation of DCs. RESULTS Inhibition of cross-presentation by activation of DCs The DCs present in mouse lymphoid organs can be classified into two main subgroups. The first are the ‘migratory’ DCs, which constitutively traffic from peripheral tissues to the lymph nodes, where they display classical markers of mature DCs (MHC class IIhi CD86hi)16. The second group comprises ‘resident’ DCs, which maintain an immature phenotype throughout their life cycle unless they encounter an activating stimulus16. Migratory DC do not traffic to the spleen, so in the steady state nearly all splenic DCs are resident immature DC17. As shown before, intravenous inoculation of the TLR ligands CpG, LPS or poly(I) poly(C) induced the maturation of splenic DCs, as assessed by upregulation of MHC class II, CD86 and CD40 (Supplementary Fig. 1 online)17–20. In our experimental conditions, none of the TLR ligands induced substantial changes in the number of DCs recovered from the spleens or lymph nodes of treated mice (data not shown). Because mature DCs are impaired in their capacity to present newly encountered antigens via MHC class II molecules21,22, we sought to determine whether such DCs were also impaired in cross-presentation. We purified DCs from the spleens of control and CpG-treated mice and tested their capacity to cross-present cell-associated ovalbumin

166

Figure 1 DCs activated in vivo downregulate cross-presentation. (a) Purified CD8 DCs from mice left untreated (control) or treated with 20 nmol CpG for 9–12 h were incubated in vitro (in duplicate) with cell-associated OVA and CFSE-labeled OT-I cells. The experiment was done once with 5 and 10 nmol CpG and multiple times with 20 nmol CpG. (b) Transgenic mice expressing act-mOVA and B6 control mice were left untreated (Untr.) or were treated with CpG; 9 h later, purified CD8 DCs from these mice were incubated in vitro (in duplicate) with CFSE-labeled OT-I cells. Results are representative of two experiments. (c) Enriched DCs from untreated and CpG-treated mice were infected in vitro with HSV, then were incubated with CFSE-labeled gBT-I cells. Results are the average of three experiments each done in duplicate; error bars indicate s.d.

(OVA) to naive OVA-specific OT-I cells in vitro (Fig. 1a). Splenic DCs can be categorized into three populations (CD4+, CD8+ and CD4– CD8– DCs), but as has been shown before, the conventional CD8 DC subset was by far the most efficient at cross-presentation2 (data not shown). Pre-exposure of splenic DCs to CpG in vivo had a dosedependent inhibitory effect on cross-presentation by CD8 DCs (Fig. 1a) and did not enhance the poor cross-presentation capacity of the non-CD8 DC types (data not shown). The inability of CpGtreated CD8 DCs to induce OT-I cell proliferation was not related to a loss of costimulatory capacity induced by the TLR agonist, as CpGtreated DCs from transgenic mice expressing OVA endogenously (Fig. 1b) or CpG-treated DCs from normal mice incubated with synthetic OVA peptide in vitro (data not shown) stimulated OT-I cell proliferation more efficiently than did their control counterparts. Presentation of endogenous viral antigens by activated DCs To test whether pretreatment with CpG also affected the endogenous or ‘classical’ MHC class I presentation pathway, we assessed presentation of an endogenous viral antigen. We isolated CD8 and non-CD8 DCs from the spleens of control and CpG-treated mice and infected them in vitro with herpes simplex virus type 1 (HSV). We then compared the infected DCs for their capacity to stimulate HSV glycoprotein B (gB)–specific gBT-I naive T cells. Infected DCs from both untreated and CpG-treated mice promoted efficient gBT-I cell proliferation (Fig. 1c), indicating that their classical MHC class I pathway was intact and confirming that both cross-presenting (CD8+) and non-cross-presenting (CD8–) DCs from CpG-injected mice could activate naive T cells. Impairment of cross-priming in mice treated with TLR ligands Although we have shown that CD8 DCs isolated from mice treated with TLR ligand were unable to cross-present cell-associated antigen in vitro, it was unclear whether this would occur in vivo, where other cell types might contribute to cross-presentation. Therefore, we injected mice with CpG, LPS or poly(I) poly(C) and, 9–12 h later, we injected the same mice with cell-associated OVA and carboxyfluorescein diacetate succinimidyl diester (CFSE)–labeled OT-I cells. We then assessed cross-presentation 60 h later by measuring the extent of OT-I cell proliferation in the spleen. Cross-presentation was impaired by pretreatment of mice with the TLR ligands (Fig. 2a). To test whether CpG caused general suppression of T cell proliferation in vivo, we assessed the response of OT-I cells in CpG-pretreated CD11c-OVA mice, which express membrane-bound OVA driven by the CD11c promoter (mostly DC specific23). In these mice, DCs constitutively present endogenously produced OVA and thus do not require cross-presentation for OT-I stimulation. OT-I cells proliferated in CD11c-OVA mice preinjected with CpG (Fig. 2b). We then tested whether OT-I proliferation could be restored in mice pretreated with

VOLUME 7

NUMBER 2

FEBRUARY 2006

NATURE IMMUNOLOGY

ARTICLES Figure 2 Activation of DCs in vivo impairs crosspriming. (a) Mice given CpG, LPS, poly(I)poly(C) CD11c:OVA B6 Untreated n = 12 100 (I:C) or left untreated for 9–12 h received cellUntreated CpG associated OVA and CFSE-labeled OT-I cells. Proliferation was assessed 60 h later with 75 representative plots (left) and quantification of proliferating cells (right). Results are 50 CpG CFSE CFSE CFSE representative of multiple experiments; numbers above bars indicate total mice analyzed. 25 n = 14 n = 4 n = 8 (b) CD11c-OVA–transgenic mice left untreated or 100 80 given CpG for 9 h received CFSE-labeled OT-I n=2 n=9 0 cells and proliferation was measured after 60 h. Untr. CpG LPS I:C CFSE 75 Nontransgenic mice were analyzed in parallel (B6). Data are representative of four experiments DC + OVA(257–264) DC no peptide 50 40 with two to three mice per group. (c) DCs Untreated CpG enriched from mice given CpG (9 h earlier) were 25 incubated in vitro with OVA(257-264) or no peptide, and then were injected into CpG-treated 0 0 or untreated mice. Results are representative Untr. 3 9 24 72 120 168 Untr. CpG injection (h) Time after CpG CFSE CFSE CFSE of two experiments with two mice per group. (d) Untreated (Untr.) or CpG-treated mice received cell-associated OVA and CFSE-labeled OT-I cells. Two mice were analyzed per time point. (e) Untreated or CpG-treated mice received cell-associated OVA and LPS; 5 d later, anti-OVA CTLs were assessed. Results are the mean of three experiments with a total of nine mice per group.

b

c

e

% specific lysis


d

CpG by injecting DCs loaded ex vivo with an OVA peptide of amino of cell-associated OVA via the MHC class II pathway to OT-II cells was acids 257–364 (OVA(257–264)). We purified DCs from CpG-treated also impaired in the mice pretreated with TLR ligand (Supplementary mice, incubated them with OVA(257–264) and injected them into Fig. 2 online). control or CpG-pretreated mice. These same mice received CFSElabeled OT-I cells in separate injections (Fig. 2c). OT-I cells prolifer- Inhibition of response to virus in mice pretreated with TLR ligands ated in both groups of mice, demonstrating their capacity to respond HSV antigens are presented in vivo by only CD8 DCs24–26, but it has in CpG-pretreated mice when their cognate antigen was presented not yet been established whether this occurs by cross-presentation. As on DCs. We obtained similar results with DCs purified from CD8 DCs pre-exposed to CpG had impaired cross-presentation, but OVA+ transgenic mice instead of DCs incubated in vitro with could still be infected in vitro with HSV and present endogenous viral OVA(257–264) (data not shown). These observations showed that antigens (Fig. 1c), we tested the effect of pretreatment with TLR ligand CpG pretreatment led to impaired cross-presentation but not to on the presentation of HSV antigens in vivo. We injected mice general immunosuppression. We then assessed the duration of the intravenously with CpG and 12 h later infected them intravenously effect of a single injection of CpG on cross-presentation. We treated with HSV. To measure presentation of HSV antigens, we also injected mice with CpG and then injected them intravenously with cell- mice with CFSE-labeled gBT-I cells at the time of infection (Fig. 4a). associated OVA and CFSE-labeled OT-I cells at different times. We Lack of proliferation of gBT-I cells in CpG-pretreated mice indicated assessed OT-I proliferation 60 h later for each time point. Down- that CD8 DCs were unable to present HSV antigens when crossregulation of cross-presentation was evident by 3 h after CpG presentation was downregulated by such pretreatment. We noted similar effects in mice pretreated with LPS or poly(I) poly(C) pretreatment and lasted for at least 5 d (Fig. 2d). To test whether CpG pretreatment impaired cross-priming of a (Fig. 4a). Furthermore, these adjuvants had similar effects on normal T cell repertoire, we treated mice with CpG and then 9–12 h responses to influenza virus in BALB/c mice (Fig. 4a), demonstrating later injected them intravenously with cell-associated OVA. We then that the effect of DC preactivation was not restricted to a particular measured their OVA-specific CTL activity in vivo 5 d later (Fig. 2e). virus, antigen or mouse strain. Pretreatment with CpG inhibited the cross-priming of CTL immunity To ensure that the effect of the TLR ligands was not due to a wider to cell-associated OVA. suppression mechanism of T cell activation, we used a ‘staggered’ Our results showed that DCs matured in vivo by TLR ligand two-antigen priming regimen in which only the HSV-specific injection were unable to cross-present cell-associated antigen and response was selectively shut down. We injected mice with CpG and therefore were incapable of inducing CTL responses by cross-priming, while maintaining their ability to induce T cell proliferation. No beads Untreated CpG LPS I:C The mechanism responsible for this effect 0.5 51 10 9 6 was most likely the downregulation of 0.2 46 12 11 22 8–10 phagocytosis that occurs in mature DCs , which would prevent access of the cellular antigen to the cross-presentation pathway. FITC Indeed, pretreatment with CpG, LPS or poly(I) poly(C) considerably downregulated Figure 3 Activation stimuli downregulate the phagocytic activity of DCs in vivo. Untreated mice or mice treated with TLR ligand were injected intravenously with fluorescent latex beads 9 h after treatment; the phagocytic activity of splenic DCs in vivo, 3 h later, splenic DCs were purified, stained with anti-CD11c and anti-CD8 and analyzed by flow as assessed by measurement of the uptake of cytometry. Numbers in quadrants indicate percentage of CD11c+CD8+ DCs (top right) and CD11c+CD8– fluorescent latex beads injected intravenously DCs (bottom right) that had captured at least one bead (fluorescein isothiocyanate (FITC)–positive (Fig. 3). In support of that view, presentation cells). Results are representative of at least two experiments. CD8



a

NATURE IMMUNOLOGY

VOLUME 7

NUMBER 2

FEBRUARY 2006

167

ARTICLES

CpG

100

n=6

Untreated

75 50 CpG n=6n=6n=9

25

75 50

+ CpG

OT-I (OVA)

gBT-I (HSV)

CFSE

c

d 6 5

CFSE

Untreated LPS CpG I:C

4 3 2

CFSE

Untr. CpG LPS I:C

n=3

100 Dividing cells (% relative to control)

Untreated

0 Untr. CpG LPS I:C

DNA copies per spleen (log10)

Untreated n=2

75 50 25 0

0 20 40 60 80 Time after infection (h)

CpG n=3 200 67 33 17 6 0 PFU (×103)

Figure 4 Impaired antiviral responses in mice pretreated with TLR ligands. (a) B6 or BALB/c mice left untreated or injected with TLR ligand received either HSV and CFSE-labeled HSV-specific gBT-I cells (B6; left) or influenza virus and influenza-specific CL4 T cells (BALB/c; right) 9 h later. Representative plots are graphed relative to values obtained in untreated mice. Numbers above bars indicate total mice analyzed. (b) Mice given cell-associated OVA without CpG (untreated) or with CpG (+ CpG) for 9–12 h received HSV, CFSE labeled-gBT-I and CFSE labeled-OT-I cells; proliferation of OT-I cells (top) or gBT-I cells (bottom) was assessed. Results are representative of two experiments with at least two mice per group. (c) Real-time PCR of viral DNA from untreated or TLR ligand–treated mice infected with HSV. Three mice were analyzed per time point. Data represent mean ± s.d. (d) Proliferation of T cells in the spleens of untreated or CpG-treated mice infected with HSV plus CFSE-labeled gBT-I cells 9–12 h after CpG treatment. Proliferation of gBT-I cells is expressed relative to that in mice infected with 200 103 PFU of HSV. Two untreated and three CpG-treated mice were analyzed per data point. Bars indicate range of values obtained.

OVA simultaneously and 9–12 h later infected these same mice with HSV and gave them CFSE-labeled OT-I and gBT-I cells. This regimen enabled us to simultaneously monitor the responses to OVA, which should not be affected by injection of CpG together with this antigen, along with the response to HSV, which would be expected to be inhibited by the shutting down of presentation induced by ‘predelivery’ of the adjuvant. We found that cell-associated OVA given at the same time as CpG was efficiently cross-presented, whereas presentation of HSV administered 9–12 h later was inhibited (Fig. 4b). Notably, inhibition of HSV cross-presentation occurred at a time when CD8 DCs were capable of priming OVA-specific T cell responses, confirming that the CpG treatment did not cause the loss of this DC population. This shows that CpG pretreatment does not

prevent CTL priming itself; instead, it only inhibits cross-presentation (formation of MHC class I complexes) of exogenous antigens subsequently encountered (more than 3 h later). To test the possibility that lack of presentation of HSV antigen in mice treated with TLR ligand could have been due to a direct effect of the innate immune response on viral load, we measured viral genomic DNA at 1, 24 and 72 h after infection in the spleens of control mice or mice pretreated with TLR ligand. Viral loads were similar in all groups of mice at the 1- and 72-hour time points and were reduced by only two thirds at 24 h in the mice treated with TLR ligand (Fig. 4c). The small difference noted at the 24-hour time point was unlikely to explain the substantial differences in T cell responsiveness, as gBT-I cell proliferation could be detected in control mice infected with as few as 2.47 103 plaque-forming units (PFU) of HSV, whereas there was no proliferation in mice pretreated with CpG that received an 80-fold higher dose of virus (Fig. 4d).

Impaired viral antigen presentation Intravenous administration of TLR ligands resulted in the activation of CD8 DCs not only in the spleen but also in the lymph node (Supplementary Fig. 1 online). We therefore determined the effect of TLR ligand pretreatment on responses initiated in the lymph node using an alternative model of HSV infection in which the virus is applied to a small area of abraded skin27. In this model, the virus replicates at the site of infection, colonizes the local nerve ganglion and then reinfects the skin innervated by the ganglion, producing lesions along a lateral band (data not shown). Injection of CpG intravenously 9 h before virus inoculation did not decrease the viral load at the primary site 48 h after infection (Fig. 5a) or the production of secondary lesions 5 d later (data not shown). However, the efficiency of presentation of HSV antigens by CD8 DCs purified from the local lymph node 2 d after infection was reduced by 80% by CpG pretreatment (Fig 5b, left). To assess if this effect correlated with impaired capacity of lymph node DC to cross-present, we also purified CD8 DCs from lymph nodes draining the contralateral (uninfected)

Figure 5 Impaired presentation of HSV antigens in the lymph node of gBT-I OT-I 40 6 30 CpG-treated mice in a zosteriform model of infection. (a) HSV in the primary site of infection 48 h after mice were inoculated with HSV on 30 20 a small area of abraded skin. Data represent results obtained in eight different mice per group, analyzed in two separate experiments; each symbol 20 5 corresponds to one mouse. Horizontal bars indicate values of the mean. 10 10 (b) Untreated or CpG-treated mice were infected in the skin with HSV 12 h after CpG inoculation; 48 h later, CD8 DCs from the lymph nodes draining 4 0 0 the infection site (left) or those from the contralateral site (right) were Untreated CpG Untreated CpG Untreated CpG purified by preparative flow cytometry (lymph nodes from 8–12 mice pooled per group). CD8 DCs from the infected site were incubated for 60 h with CFSE-labeled gBT-I cells for measurement of presentation of viral antigen (left); those from the contralateral site were incubated with cell-associated OVA and CFSE-labeled OT-I cells for assessment of their capacity to cross-present OVA (right). Total proliferating T cells were determined by flow cytometry 60 h later. Each measurement was done in duplicate; each symbol represents the result of one experiment of a total of four; horizontal bars represent values of the mean.

168

VOLUME 7

Dividing cells (×103)

b

Dividing cells (×103)

a

HSV PFU (log10)


b

n = 4 n = 4n = 4

25

0 CFSE

n=5

100 Dividing cells (% relative to control)

Untreated Dividing cells (% relative to control)

a

NUMBER 2

FEBRUARY 2006

NATURE IMMUNOLOGY

ARTICLES

25

0

Untreated

CpG

Figure 6 Impairment of endogenous anti-HSV CTL responses in CpG-treated mice. Untreated or CpG-treated mice were injected intravenously with HSV; 5 d later, induction of an endogenous CTL response was determined by in vivo CTL assay. Results are representative of two experiments with a total of six mice per group. Error bars show range of values obtained.

flanks of the same mice and assessed their capacity to cross-present OVA in vitro (Fig. 5b, right). The results showed a 76% reduction in cross-presenting capacity, providing a notable correlation between the ability of CpG pretreatment to impair presentation of cell-associated OVA and its ability to impair presentation of viral antigens. No generation of anti-HSV CTL responses in CpG-treated mice Next we determined whether CpG pretreatment affected the priming of CTLs from the endogenous repertoire. We pretreated mice with CpG and the next day infected them intravenously with HSV. We then assessed in vivo CTL activity 5 d later. We detected HSV-specific CTL activity in normal mice, but it was greatly diminished in mice pretreated with CpG (Fig. 6). This result demonstrated that systemic administration of an activating stimulus can impair the induction of endogenous CTLs against subsequent viral infections. Impairment of cross-presentation by malaria parasite infection Intravenous injection of TLR ligands probably mimics the effects of sepsis or systemic parasitic infection. To directly address that possibility, we measured cross-presentation after malaria parasite infection, reported to activate DCs by TLR9 (refs. 28,29). Intravenous injection of blood-stage forms of the murine malaria pathogen Plasmodium berghei (ANKA strain) produced a systemic infection of the red blood cells that led to a gradual increase in parasitemia (Fig. 7a). Examination of the phenotype of DCs during the course of infection showed increased expression of MHC class II and costimulatory molecules and downregulation of phagocytosis at days 3 and 4 after infection, indicative of systemic DC activation analogous to that caused by injection of TLR ligand (Fig. 7b and data not shown). Such DC activation was correlated with a reduction in the capacity of isolated CD8 DCs to cross-present cell-associated OVA

DISCUSSION As immature cells, CD8 DCs are constitutively equipped to crosspresent soluble and cellular antigens2. Here we have shown that mature DCs downregulated their capacity to cross-present newly encountered antigens in vivo30 while maintaining their ability to present endogenous antigens such as transgenic OVA or, when infected with HSV, viral antigens. The mechanism responsible for the lack of cross-presentation by mature DCs was most likely the downregulation of phagocytosis that accompanies DC maturation8–10, although additional impairment in the cross-presenting pathway ‘downstream’ of antigen uptake was also likely to have contributed (unpublished results). Downregulation of cross-presentation was a physiological response to DC maturation: mice preinjected with LPS, CpG or

a

NATURE IMMUNOLOGY

VOLUME 7

NUMBER 2

FEBRUARY 2006

Uninfected LPS P. berghei

10 8 6

MHCII

4 2 0

Figure 7 P. berghei infection causes systemic DC activation and downregulates cross-presentation. (a) Time course of parasitemia after infection with 1 106 P. berghei parasites. Data represent mean ± s.e.m. for four mice analyzed in each of two experiments. (b) Surface expression of MHC class II (MHCII) and CD86 in splenic CD8 DCs from uninfected, LPS-treated (24 h before) or P. berghei-infected (3 d before) mice. Backgr., background fluorescence of unstained cells. Similar results were obtained for non-CD8 DCs (data not shown). Results are representative of three experiments. (c) Proliferation of OT-I cells in the presence of splenic CD8 DCs from infected or uninfected mice incubated with cell-associated OVA (left) or OVA(257-264) (right). Proliferation was assessed in duplicate and is presented relative to that in uninfected mice. Data represent mean ± s.e.m. obtained in two experiments.

Backgr.

b

12

0

1

2

3

4

5

6

7

Time after infection (d)

c

CD86 300

100


50

% red blood cells infected

75

in vitro (Fig. 7c, left) even though these DCs were more potent in inducing T cell proliferation when loaded with synthetic OVA peptide (Fig. 7c, right). To assess the effect of malaria parasite infection on crosspresentation in vivo, we injected mice with cell-associated OVA and CFSE-labeled OT-I cells at various times after parasite inoculation and measured proliferation in the spleen 60 h later. OT-I cell proliferation was ‘boosted’ 1 d after infection, suggesting an initial adjuvant effect of the parasite on cross-presentation of OVA. However, cross-presentation, as indicated by OT-I cell responses, decreased in mice injected on day 2 and was considerably impaired in mice injected 3 or 4 d after infection (Fig. 8a), time points at which maturation of the DC was evident (Fig. 7b). As was true in mice pretreated with TLR ligands, naive OT-I cells were able to proliferate in malaria-infected CD11c-OVA mice, whose DCs constitutively present endogenous OVA (Fig. 8b). Moreover, proliferation of OT-I cells was induced in malaria-infected mice by injection of DCs loaded with peptide antigen (Fig. 8c). Although such proliferation was not as efficient as that in uninfected mice, indicating the involvement of additional inhibitory components, these findings clearly suggest an avenue for inducing immunity in an environment in which endogenous DCs are activated and unable to cross-prime.



Specific lysis (%)

100

75

50

25

0

Uninfected

Day 3 Day 4 P. berghei

0.41 nM

1.23 nM

250 200 150 100 50 0

Uninfected

Day 3 Day 4 P. berghei

169

ARTICLES

a


b

Uninfected

Day 1

Day 2

Day 3

Day 4

CFSE

CFSE

CFSE

CFSE

CFSE

CD11c-OVA Uninfected P. berghei

B6

CFSE

c

CFSE

DC + OVA(257–264) Uninfected P. berghei

CFSE

CFSE

CFSE DC no peptide

CFSE

Figure 8 Impaired cross-presentation in P. berghei-infected mice. (a) OT-I cell proliferation in the spleen in response to cell-associated OVA injected into mice infected for 1–4 d with P. berghei. Data are representative of two independent experiments, with a total of four mice analyzed per time point. (b) CD11c-OVA mice left uninfected or infected with P. berghei for 3 d were injected with CFSE-labeled OT-I cells and proliferation was measured after 60 h. Nontransgenic mice were analyzed in parallel (B6). Results are representative of three experiments, each with two or three mice per group. (c) Enriched DC preparations from CpG-treated mice were incubated in vitro with OVA(257–264) or no peptide, and then were injected into uninfected mice or mice infected with P. berghei for 3 d. All mice then received CFSElabeled OT-I cells. Results are representative of two experiments, with a total of four mice per group.

poly(I) poly(C) or undergoing malaria parasite infection, situations that induced systemic DC maturation, lacked cells capable of crosspresenting cell-associated OVA and could not generate OVA-specific CTLs by cross-priming. Although direct encounter with TLR ligands is an obvious route to DC maturation, it is possible that some maturation was caused indirectly by cytokines released from TLR ligand– responsive cells31. This is particularly likely for poly(I) poly(C), which caused maturation of all splenic DCs despite the lack of TLR3 in some DC subtypes20,32 and their inability to respond directly to poly(I) poly(C)33. Several reports have suggested that malaria parasites of both human and rodent origin can impair DC function by blocking activation of DCs34–36. However, consistent with our findings, mouse DCs have been found to be effectively activated after infection with various plasmodium species37–39. A full explanation for these discrepancies will require further investigation, but this probably involves differences in parasite species and the time at which DCs were analyzed after infection, both of which varied considerably from study to study. Notably, preactivation of DCs in vivo with TLR ligands substantially impaired the presentation of HSV and influenza viral antigens. This effect was independent of the mouse strain or route of infection (intravenous or cutaneous) and could not be explained by decreased viral infectivity in the mice treated with TLR ligand, because the viral load decreased little (intravenous model) or not at all (cutaneous model), and the replication cycle proceeded normally. Because in vitro infection of CD8 DCs from mice treated with TLR ligand allowed efficient viral antigen presentation, the simplest interpretation for impaired immunity to viruses in vivo is that in this circumstance, the DCs are not infected but instead rely heavily on cross-presentation to prime CTLs. These observations provide strong support for the

170

debated hypothesis that CTL immunity to viral infections can be induced by cross-priming3–6. Several observations excluded the possibility that TLR ligand pretreatment or malaria parasite infection stimulates DCs to either differentiate to a state unable to activate T cells in vivo or to induce suppressive factors or regulatory cells that prevent T cell proliferation. First, OT-I cells were able to proliferate in CpG-treated or malariainfected CD11c-OVA mice, in which DCs present OVA via the endogenous pathway rather than by cross-presentation. Second, inoculation of mature DCs bearing OVA peptide into mice pretreated with CpG or infected with P. berghei induced OT-I proliferation. Third, administration of cell-associated OVA at the time of CpG injection led to normal responses to this antigen at a time when responses to a subsequent challenge with HSV (in the same mouse) were inhibited. That result demonstrated that DCs exposed to CpG and OVA simultaneously were alive and presenting antigen, whereas the same DCs failed to present HSV antigen encountered subsequently. T cells can thus proliferate efficiently in mice pretreated with TLR ligands or undergoing malaria infection, provided they encounter DCs presenting cognate MHC class I–peptide complexes, at least in the experimental models we assessed here. Our study raises the issue of why cross-presentation is downregulated after DC maturation. The likely answer is that this regulation enables mature CD8 DCs to ‘focus’ their cross-priming activity on antigens associated with the signal that triggered their activation, a situation analogous to that described before for the MHC class II presentation pathway1,21,40,41. Such downregulation, therefore, should not normally be deleterious in most physiological situations in which maturation probably only affects a limited number of DCs at a particular location. However, there are several situations in which our findings may have important implications. Vaccination formulations should provide DC activation signals in close physical and temporal association with the antigen to prevent premature inactivation of the DC. Moreover, treatments designed to induce systemic DC activation in vivo should include monitoring of their potential immunosuppressive effect11. Our results also help to explain the immunosuppressive effects of sepsis14 and malaria12,13,15. Although it is likely that multiple mechanisms are involved in these cases, our study has indicated that downregulation of cross-presentation is an important consequence of systemic DC activation. That consequence raises some concerns regarding the use of live viral vectors for vaccination programs in malariaendemic areas, which may be limited by the extent of cross-presenting DCs available in parasite-infected people. Nevertheless, the finding that provision of DCs presenting the right MHC class I–peptide combination allowed the activation of OT-I cells in TLRpretreated or malaria parasite–infected mice suggests that vaccination protocols that bypass the requirement for cross-presentation in vivo should provide a useful strategy for overcoming the effects of systemic blood infection. METHODS Mice. The mice used were C57BL/6 (B6), B6.C-H-2bm-1 (bm1) and BALB/c (I-Ad) and the transgenic strains act-mOVA42, OT-I (ref. 43), CL4 (ref. 44), gBT-I (ref. 45), OT-II (ref. 46) and CD11c-OVA. The CD11c-OVA strain expresses OVA under control of the CD11c promoter23 (R.J.S., unpublished data). Where indicated, mice were injected intravenously in the tail vein with 5–20 nmol of synthetic phosphorothioated CpG1668 (GeneWorks), 25–100 mg poly(I) poly(C) (Sigma) or 3 mg LPS (Sigma) dissolved in PBS. All mice were maintained in specific pathogen–free conditions in the animal facilities of the Walter and Eliza Hall Institute (Parkville, Australia) following institutional guidelines and were used between five and 12 weeks of age.

VOLUME 7

NUMBER 2

FEBRUARY 2006

NATURE IMMUNOLOGY

ARTICLES


DC isolation, analysis and culture. DC purifications from spleen or lymph node, analytical and preparative flow cytometry and DC cultures in vitro were done as described17,47,48. Preparation of CFSE-labeled T cells. OT-I T cells (H-2Kb-restricted antiOVA(257–264), gBT-I T cells (H-2Kb-restricted anti-HSV gB(498–505), CL4 T cells (H-2Kd-restricted anti-influenza hemagglutinin(512–520)) or OT-II T cells (I-Ab-restricted anti-OVA(323–339)) were purified from pooled lymph nodes (inguinal, axillary, sacral, cervical and mesenteric) of transgenic mice by depletion of non-CD8 T cells (non-CD4 T cells for purification of OT-II samples) and were labeled with CFSE as described17. The T cell preparations were routinely 85–95% pure, as determined by flow cytometry. Preparation of cell-associated OVA. Spleen cells from bm1 mice were irradiated with 1,500 cGy, were washed, were incubated for 10 min at 37 1C with 10 mg/ml of OVA in FCS (Sigma) and RPMI 1640 medium (Sigma) and were washed three times in RPMI 1640 medium containing 2.5% (volume/ volume) FCS. The amount of OVA associated with the cells was approximately 8 ng OVA per 1 106 cells49. Quantitation of OT-I cell proliferation in vitro. Purified CD8 DCs from untreated or CpG-pretreated act-mOVA mice were plated in V-bottomed 96-well plates (Costar) with 5 104 CFSE-labeled transgenic OT-I cells. Proliferation was quantified after 60 h of culture as described17. OT-I cells were labeled with antibody to CD8 (anti-CD8) and anti-Va2 and were resuspended in 100 ml balanced-salt solution and 3% (volume/volume) FCS containing 3 104 blank calibration particles (BD Biosciences Pharmingen). Samples were analyzed by flow cytometry on an LSR (Beckton Dickinson) and the total number of live dividing lymphocytes (propidium iodide–negative, CFSElo) was calculated from the number of dividing cells per 5 103 beads. Each determination was done in duplicate. For experiments with cell-associated OVA, 2.5 104 purified CD8 DCs were plated in U-bottomed 96-well plates (Costar) with 2.5 105 OVA-coated bm1 splenocytes and 5 104 CFSElabeled OT-I cells and were incubated for 60 h. The number of proliferating T cells was determined as described above. For experiments on the presentation of OVA(257–264), CD8 DCs purified from uninfected or malaria-infected mice were incubated for 45 min in V-bottomed 96-well plates (Costar) with various concentrations of OVA(257–264) and were washed; 5 104 CFSE-labeled OT-I cells were added to each well. The cells were incubated for 60 h and proliferation of OT-I cells was assessed as described above. In vitro infection with HSV. DC-enriched preparations, 70–80% pure, were washed three times in RPMI 1640 medium and were incubated for 45 min at 37 1C with HSV (KOS strain) at a dose of 3 PFU/cell at a final concentration of 2 106 cells/ml. The DCs were washed three times with EDTA and balanced-salt solution supplemented with 10% FCS, were stained and were sorted into CD8 or non-CD8 DCs. Each population was incubated for 60 h with 5 104 CFSE-labeled transgenic gBT-I T cells. The gBT-I cells were then labeled with anti-Va2 and anti-CD8 and proliferation was measured as described above. Cross-presentation and MHC class II presentation of OVA in vivo. Untreated, TLR ligand–treated or malaria-infected mice were injected intravenously with 2 107 OVA-coated spleen cells and 2 106 to 3 106 CFSE-labeled OT-I cells or OT-II cells. After 60 h, splenocyte suspensions that had been depleted of red blood cells were stained with anti-Va2 plus anti-CD8 (OT-I) or anti-CD4 (OT-II). Samples were washed and resuspended in 200 ml balanced-salt solution plus 3% (volume/volume) FCS, then were analyzed by flow cytometry. Presentation of viral antigens in the intravenous model of infection. Untreated mice or mice treated with TLR ligand were infected intravenously with 2 105 PFU HSV (or threefold dilutions thereof) or 8.8 104 PFU influenza A/PR/8/34 (H1/N1) (Mount Sinai strain; a gift from L. Brown, University of Melbourne, Melbourne, Australia) diluted in PBS and were injected intravenously separately with 1 106 to 2 106 CFSE-labeled gBT-I or CL4 cells, respectively. Then, 60 h later, mice were killed and gBT-I or CL4 proliferation in the spleen was determined by flow cytometry.

NATURE IMMUNOLOGY

VOLUME 7

NUMBER 2

FEBRUARY 2006

Presentation of HSV antigens in the flank infection model. Mice were inoculated by flank scarification as described27. CD8 DCs were purified from brachial lymph nodes draining the infected and uninfected sides of control or CpG-pretreated mice and were incubated in vitro with 5 104 CFSE-labeled gBT-I cells or with OVA-coated splenocytes and CFSE-labeled OT I cells as described above. T cell proliferation was determined by flow cytometry 60 h later. In vivo CTL assay. Suspensions of a mixture of splenocytes and lymph node cells from B6 mice were depleted of red cells and were split into two equal portions. One was pulsed for 1 h at 37 1C with either 0.1 mg/ml of OVA(257–264) or 0.01 mg/ml of gB(498–505) and then was labeled with a high concentration (2.5 mM) of CFSE. The other was incubated for 1 h at 37 1C without peptide and was labeled with a low concentration (0.25 mM) of CFSE (CFSElo population). Equal numbers of cells from each population were combined and 2 107 cells in total were adoptively transferred by intravenous injection into mice 5 d after priming with HSV or with OVA-coated bm1 splenocytes together with 1 mg LPS. Mice were killed 4 h later. Spleen cell suspensions were analyzed by flow cytometry and each population was distinguished by fluorescence intensity. Percent OVA- or gB-specific lysis was determined by loss of the peptide-pulsed CFSEhi population compared with the control CFSElo population. Quantification of HSV in the skin and spleen. The amount of lytic virus in the skin was determined as described27. The amount of lytic virus in the spleens of mice infected intravenously was below the detection limit even in control mice, so a real-time PCR method to measure the amount of genomic DNA, adapted from that described50, was used. TaqMan real-time PCR amplification and detection reactions were done with an ABI 7700 Sequence Detector (Applied Biosytems) with HSV thymidine kinase gene–specific primers (5¢-TTGTCTCCTTCCGTGTTTCAGTT-3¢ and 5¢-GGCTCCATACCGACGAT CTG-3¢) and a fluorescence-labeled probe (5¢–6-carboxyfluorescein–CCA TCTCCCGGGCAAACGTGC–N,N,N¢,N¢-tetramethyl-6-carboxyrhodamine–3¢) for detection of HSV viral DNA. Reactions were done in 25-ml volumes containing Platinum Quantitative PCR Supermix-UDG and ROX reference dye (Invtirogen) with a final concentration of 200 nM of each primer and 100 nM TaqMan probe. PCR reactions consisted of 50 1C for 2 min and 95 1C for 10 min, then 55 cycles of 95 1C for 30 s (denaturation) and 60 1C for 1 s (annealing and extension), followed by storage at 4 1C. Total genomic DNA was extracted from 10-mg spleen samples with the QIAGEN Dneasy kit or QIAamp DNA Micro kit (Qiagen). Duplicates of each sample DNA were analyzed in parallel with duplicates of viral DNA standards to determine the quantity of viral DNA copies. For viral DNA standards, purified HSV KOS strain viral DNA was serially diluted such that 10 ml of the sample contained 1 106, 1 105, 1 104, 1 103, 1 102, 1 101 or 1 100 copies of HSV KOS strain viral DNA. As little as one viral DNA copy could be detected routinely in these assays. In the experiment presented, duplicate quantification was made on triplicate samples and results were normalized on a perspleen basis. Malaria infections. P. berghei ANKA strain parasites were used. These were essentially wild-type except that the line used constitutively expresses green fluorescent protein51. Mice were infected intravenously with 1 106 P. berghei parasites. Parasitemia was monitored by examination of Giemsa-stained smears of blood obtained from the tail vein. Injection of DCs loaded with OVA peptide into malaria-infected or CpGtreated mice. DCs were purified from CpG-injected B6 mice 9 h after injection, were incubated with 1 mg/ml of synthetic OVA(257–264) (Mimetopes) and were washed. Control, TLR ligand–treated or malaria-infected mice were injected separately in the opposite tail veins with 2 106 OVA(257–264)– loaded DCs and 2 106 CFSE-labeled OT-I cells. Phagocytosis of fluorescent microspheres. Mice were injected intravenously with 3.64 1010 Fluoresbrite YG carboxylate microspheres (0.5 mm; Polysciences). Then, 3 h later, spleens were collected and DCs were purified as described above. Uptake was assessed 3 h later by flow cytometry. Note: Supplementary information is available on the Nature Immunology website.

171

ARTICLES


ACKNOWLEDGMENTS We thank L. Buckingham, K. Gray, D. John, C. Jordan, F. Kupresanin, J. Langley, J.-L. Tan and all members of the Flow Cytometry and the Animal Services facilities at the Walter and Eliza Hall Institute for technical assistance. Supported by the National Health and Medical Research Council of Australia (G.T.B., F.R.C., K.S., B.S.C., W.R.H. and J.A.V.), the Anti-Cancer Council of Australia (J.A.V.), the Deutsche Forschungsgemeinschaft BE 2089/1-1 (G.M.N.B.), the Wellcome Trust (G.T.B.), Howard Hughes Medical Institute (G.T.B., B.S.C. and W.R.H.) and the Leukemia and Lymphoma Society (J.A.V.). COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. Published online at http://www.nature.com/natureimmunology/ Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/

1. Wilson, N.S. & Villadangos, J.A. Regulation of antigen presentation and crosspresentation in the dendritic cell network: facts, hypothesis, and immunological implications. Adv. Immunol. 86, 241–305 (2005). 2. Heath, W.R. et al. Cross-presentation dendritic cell subsets, and the generation of immunity to cellular antigens. Immunol. Rev. 199, 9–26 (2004). 3. Sigal, L.J., Crotty, S., Andino, R. & Rock, K.L. Cytotoxic T-cell immunity to virusinfected non-haematopoietic cells requires presentation of exogenous antigen. Nature 398, 77–80 (1999). 4. Melief, C.J. Mini-review: Regulation of cytotoxic T lymphocyte responses by dendritic cells: peaceful coexistence of cross-priming and direct priming? Eur. J. Immunol. 33, 2645–2654 (2003). 5. Freigang, S., Egger, D., Bienz, K., Hengartner, H. & Zinkernagel, R.M. Endogenous neosynthesis vs. cross-presentation of viral antigens for cytotoxic T cell priming. Proc. Natl. Acad. Sci. USA 100, 13477–13482 (2003). 6. Norbury, C.C. & Sigal, L.J. Cross priming or direct priming: is that really the question? Curr. Opin. Immunol. 15, 82–88 (2003). 7. Medzhitov, R. Toll-like receptors and innate immunity. Nat. Rev. Immunol. 1, 135–145 (2001). 8. Garrett, W.S. et al. Developmental control of endocytosis in dendritic cells by Cdc42. Cell 102, 325–334 (2000). 9. West, M.A., Prescott, A.R., Eskelinen, E.L., Ridley, A.J. & Watts, C. Rac is required for constitutive macropinocytosis by dendritic cells but does not control its downregulation. Curr. Biol. 10, 839–848 (2000). 10. West, M.A. et al. Enhanced dendritic cell antigen capture via Toll-like receptor-induced actin remodeling. Science 305, 1153–1157 (2004). 11. Klinman, D.M. Immunotherapeutic uses of CpG oligodeoxynucleotides. Nat. Rev. Immunol. 4, 249–258 (2004). 12. Williamson, W.A. & Greenwood, B.M. Impairment of the immune response to vaccination after acute malaria. Lancet 1, 1328–1329 (1978). 13. Whittle, H.C. et al. T-cell control of Epstein-Barr virus-infected B cells is lost during P. falciparum malaria. Nature 312, 449–450 (1984). 14. Cohen, J. The immunopathogenesis of sepsis. Nature 420, 885–891 (2002). 15. Urban, B.C. & Roberts, D.J. Inhibition of T cell function during malaria: implications for immunology and vaccinology. J. Exp. Med. 197, 137–141 (2003). 16. Villadangos, J.A. & Heath, W.R. Life cycle, migration and antigen presenting functions of spleen and lymph node dendritic cells: limitations of the Langerhans cells paradigm. Semin. Immunol. 17, 262–272 (2005). 17. Wilson, N.S. et al. Most lymphoid organ dendritic cell types are phenotypically and functionally immature. Blood 102, 2187–2194 (2003). 18. De Smedt, T. et al. Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J. Exp. Med. 184, 1413–1424 (1996). 19. Kamath, A.T. et al. The development, maturation, and turnover rate of mouse spleen dendritic cell populations. J. Immunol. 165, 6762–6770 (2000). 20. Edwards, A.D. et al. Toll-like receptor expression in murine DC subsets: lack of TLR7 expression by CD8a+ DC correlates with unresponsiveness to imidazoquinolines. Eur. J. Immunol. 33, 827–833 (2003). 21. Wilson, N.S., El-Sukkari, D. & Villadangos, J.A. Dendritic cells constitutively present self antigens in their immature state in vivo, and regulate antigen presentation by controlling the rates of MHC class II synthesis and endocytosis. Blood 103, 2187–2195 (2004). 22. Villadangos, J.A., Schnorrer, P. & Wilson, N.S. Control of MHC class II antigen presentation in dendritic cells: a balance between creative and destructive forces. Immunol. Rev. 207, 191–205 (2005).

172

23. Brocker, T., Riedinger, M. & Karjalainen, K. Targeted expression of major histocompatibility complex (MHC) class II molecules demonstrates that dendritic cells can induce negative but not positive selection of thymocytes in vivo. J. Exp. Med. 185, 541–550 (1997). 24. Smith, C.M. et al. Cutting Edge: Conventional CD8a+ dendritic cells are preferentially involved in CTL priming after footpad Infection with herpes simplex virus-1. J. Immunol. 170, 4437–4440 (2003). 25. Allan, R.S. et al. Epidermal viral immunity induced by CD8a+ dendritic cells but not by Langerhans cells. Science 301, 1925–1928 (2003). 26. Belz, G.T. et al. Cutting edge: conventional CD8a+ dendritic cells are generally involved in priming CTL immunity to viruses. J. Immunol. 172, 1996–2000 (2004). 27. van Lint, A. et al. Herpes simplex virus-specific CD8+ T cells can clear established lytic infections from skin and nerves and can partially limit the early spread of virus after cutaneous inoculation. J. Immunol. 172, 392–397 (2004). 28. Pichyangkul, S. et al. Malaria blood stage parasites activate human plasmacytoid dendritic cells and murine dendritic cells through a Toll-like receptor 9-dependent pathway. J. Immunol. 172, 4926–4933 (2004). 29. Coban, C. et al. Toll-like receptor 9 mediates innate immune activation by the malaria pigment hemozoin. J. Exp. Med. 201, 19–25 (2005). 30. Gil-Torregrosa, B.C. et al. Control of cross-presentation during dendritic cell maturation. Eur. J. Immunol. 34, 398–407 (2004). 31. Sporri, R. & Reis c Sousa, C. Inflammatory mediators are insufficient for full dendritic cell activation and promote expansion of CD4+ T cell populations lacking helper function. Nat. Immunol. 6, 163–170 (2005). 32. Naik, S.H. et al. Cutting edge: generation of splenic CD8+ and CD8– dendritic cell equivalents in Fms-like tyrosine kinase 3 ligand bone marrow cultures. J. Immunol. 174, 6592–6597 (2005). 33. Schulz, O. et al. Toll-like receptor 3 promotes cross-priming to virus-infected cells. Nature 433, 887–892 (2005). 34. Urban, B.C. et al. Plasmodium falciparum-infected erythrocytes modulate the maturation of dendritic cells. Nature 400, 73–77 (1999). 35. Urban, B.C., Willcox, N. & Roberts, D.J. A role for CD36 in the regulation of dendritic cell function. Proc. Natl. Acad. Sci. USA 98, 8750–8755 (2001). 36. Ocana-Morgner, C., Mota, M.M. & Rodriguez, A. Malaria blood stage suppression of liver stage immunity by dendritic cells. J. Exp. Med. 197, 143–151 (2003). 37. Seixas, E., Cross, C., Quin, S. & Langhorne, J. Direct activation of dendritic cells by the malaria parasite, Plasmodium chabaudi chabaudi. Eur. J. Immunol. 31, 2970–2978 (2001). 38. Luyendyk, J., Olivas, O.R., Ginger, L.A. & Avery, A.C. Antigen-presenting cell function during Plasmodium yoelii infection. Infect. Immun. 70, 2941–2949 (2002). 39. Leisewitz, A.L. et al. Response of the splenic dendritic cell population to malaria infection. Infect. Immun. 72, 4233–4239 (2004). 40. Cella, M., Engering, A., Pinet, V., Pieters, J. & Lanzavecchia, A. Inflammatory stimuli induce accumulation of MHC class II complexes on dendritic cells. Nature 388, 782–787 (1997). 41. Pierre, P. et al. Developmental regulation of MHC class II transport in mouse dendritic cells. Nature 388, 787–792 (1997). 42. Ehst, B.D., Ingulli, E. & Jenkins, M.K. Development of a novel transgenic mouse for the study of interactions between CD4 and CD8 T cells during graft rejection. Am. J. Transplant. 3, 1355–1362 (2003). 43. Hogquist, K.A. et al. T cell receptor antagonist peptides induce positive selection. Cell 76, 17–27 (1994). 44. Morgan, D.J. et al. CD8+ T cell-mediated spontaneous diabetes in neonatal mice. J. Immunol. 157, 978–983 (1996). 45. Mueller, S.N., Heath, W., McLain, J.D., Carbone, F.R. & Jones, C.M. Characterization of two TCR transgenic mouse lines specific for herpes simplex virus. Immunol. Cell Biol. 80, 156–163 (2002). 46. Barnden, M.J., Allison, J., Heath, W.R. & Carbone, F.R. Defective TCR expression in transgenic mice constructed using cDNA-based a- and b-chain genes under the control of heterologous regulatory elements. Immunol. Cell Biol. 76, 34–40 (1998). 47. Vremec, D., Pooley, J., Hochrein, H., Wu, L. & Shortman, K. CD4 and CD8 expression by dendritic cell subtypes in mouse thymus and spleen. J. Immunol. 164, 2978–2986 (2000). 48. O’Keeffe, M. et al. Mouse plasmacytoid cells: long-lived cells, heterogeneous in surface phenotype and function, that differentiate into CD8+ dendritic cells only after microbial stimulus. J. Exp. Med. 196, 1307–1319 (2002). 49. Li, M. et al. Cell-associated ovalbumin is cross-presented much more efficiently than soluble ovalbumin in vivo. J. Immunol. 166, 6099–6103 (2001). 50. Zhao, X. et al. Vaginal submucosal dendritic cells, but not Langerhans cells, induce protective Th1 responses to herpes simplex virus-2. J. Exp. Med. 197, 153–162 (2003). 51. Franke-Fayard, B. et al. A Plasmodium berghei reference line that constitutively expresses GFP at a high level throughout the complete life cycle. Mol. Biochem. Parasitol. 137, 23–33 (2004).

VOLUME 7

NUMBER 2

FEBRUARY 2006

NATURE IMMUNOLOGY