In this study, we report the dynamic changes in activation and functions that occur in spleen dendritic cell (sDC) subsets following infection of mice with a natural murine pathogen, lymphocytic choriomeningitis virus (LCMV). Within 24 h postinfection (pi), sDCs acquired the ability to stimulate naive LCMV-specific CD8+ T cells ex vivo. Conventional (CD11chigh CD8+ and CD4+) sDC subsets rapidly up-regulated expression of costimulatory molecules and began to produce proinflammatory cytokines. Their tendency to undergo apoptosis ex vivo simultaneously increased, and in vivo the number of conventional DCs in the spleen decreased markedly, dropping ∼2-fold by day 3 pi. Conversely, the number of plasmacytoid (CD11clowB220+) DCs in the spleen increased, so that they constituted almost 40% of sDCs by day 3 pi. Type 1 IFN production was up-regulated in plasmacytoid DCs by 24 h pi. Analysis of DC activation and maturation in mice unable to respond to type 1 IFNs implicated these cytokines in driving infection-associated phenotypic activation of conventional DCs and their enhanced tendency to undergo apoptosis, but also indicated the existence of type 1 IFN-independent pathways for the functional maturation of DCs during LCMV infection.
Dendritic cells (DC)3 were initially identified as potent APCs with the capacity to mediate the activation of naive Ag-specific T cells (1). It is now recognized that they have a central immunoregulatory role both in the adaptive response, where they are involved in the maintenance of peripheral tolerance in addition to the triggering of Ag-specific T cell responses, and also in the innate response. In the spleen of mice, three subpopulations of mature CD11c+, MHC class II+ DCs can be distinguished on the basis of differential expression of CD4 and CD8α homodimers: CD4+CD8α−, CD4−CD8α+, and CD4−CD8α− (referred to as CD4+, CD8+, and double negative) (2). These conventional CD11chigh murine spleen DC (sDC) subsets efficiently present Ag to and stimulate the proliferation of T cells, and it has been shown that both CD8+ and CD8− sDCs are capable of inducing the priming of CD4+ (3) and CD8+ (4) T cell responses in vivo. However, only the CD8+ DC subset is able to mediate cross-priming of CD8+ T cell responses to exogenous Ags (5). CD8+ and CD8− sDC populations also differ in the cytokines they produce. Notably, CD8+ DCs produce much higher levels of IL-12p70 than CD8− DCs under many experimental conditions, and hence have a greater tendency to induce a Th1-biased cytokine profile in responding CD4+ T cells (3, 6). A fourth subpopulation of murine sDCs, the plasmacytoid DC (pDC) subset, has also been identified (7, 8), phenotypically characterized as being CD11clow, MHC class IIlow, and B220+. Although pDCs can stimulate T cell proliferation when suitably activated (9), they are relatively poor APCs. However, they have the capacity to secrete high levels of type 1 IFN in response to stimulation with viruses or bacterial DNA, and act as a major source of type 1 IFN production during many virus infections (8, 10).
DCs are highly responsive to stimuli such as microbial infection, inflammation, and tissue damage, and their maturation is further driven by interaction with T cells (11). To increase understanding of the activation and functional maturation of DC subsets stimulated by different infections and how this in turn shapes the ensuing immune response, we undertook to characterize the DC response in the early stages of acute infection of mice with lymphocytic choriomeningitis virus (LCMV). LCMV is a noncytopathic virus that can cause either acute or persistent infections in mice, the outcome of infection being dependent on both host and viral factors (12). The i.p. inoculation of adult immunocompetent mice with the Armstrong strain of LCMV results in a well-characterized acute infection during which the virus replicates systemically in tissues including the spleen, stimulating the induction of a high-magnitude LCMV-specific CD8+ CTL response (13), which plays a key role in effecting viral clearance.
In this study, we addressed the dynamic changes in the activation state and functional capacity of sDC subpopulations that pave the way for expansion of the strongly type 1-biased adaptive response in LCMV Armstrong-infected mice. Our results show that within 24 h after infection of mice with LCMV, sDCs start to produce cytokines including type 1 IFNs, TNF-α, and IL-12, and acquire potent T cell stimulatory properties. Furthermore, they indicate that optimal conditions for T cell priming are maintained for only a short period in vivo, with cytokine production by all sDC subsets being rapidly down-regulated and the number of CD11chighCD8+ and CD4+ sDCs quickly starting to decline, the latter process correlating with these cells exhibiting an enhanced apoptotic death rate ex vivo.
Importantly, we also addressed the role played by type 1 IFNs (innate cytokines that have both direct antiviral effects and an immunoregulatory role (14)), in mediating the changes we observed in the activation state, T cell stimulatory capacity, and number of conventional DCs in the spleen during this infection. Type 1 IFNs are produced early after LCMV infection and reach peak titers in the serum 2–3 days postinfection (pi) (15). Interestingly, pDCs are not thought to constitute the major source of serum type 1 IFN in LCMV-infected mice (10). Previous studies have shown that the type 1 IFN response has an important impact on the ability of mice to control LCMV replication (16). Type 1 IFNs are capable of activating conventional DCs both in vitro and in vivo (17, 18) and have potent adjuvant properties, stimulating cross-priming of Ag-specific CD8+ T cell responses (19). In this study, we show that type 1 IFNs contribute to the phenotypic activation of sDCs in LCMV-infected mice and are involved in mediating the dramatic decrease in the number of conventional DCs in the spleen that accompanies this infection. However, we demonstrate that the functional maturation of DCs to drive naive CD8+ T cell stimulation can occur by type 1 IFN-independent mechanisms.
Materials and Methods
C57BL/6 mice were obtained from Charles River Laboratories or from the specific pathogen-free unit at Institute for Animal Health. OTI transgenic mice (originally obtained from L. Lefrancois, University of Connecticut, Farmington, CT) that have a transgenic TCR specific for the OVA257–264 (SIINFEKL) epitope plus H2-Kb (20), LCMV TCR transgenic mice (originally purchased from The Jackson Laboratory) expressing a TCR specific for the GP-133–41 (KAVYNFATM) epitope plus H2-Db (21), and 129SvEv (129) and type 1 IFN receptor knockout (IFN-R KO) mice (22) (originally purchased from B&K Universal) were all obtained from the specific pathogen-free unit at Institute for Animal Health. Mice were used at 6–12 wk of age. All animal studies were conducted in accordance with United Kingdom Home Office regulations, and were approved by the site ethical review committee.
Virus growth, titration, and use for infection of mice
Spleen DCs were obtained using a variation of the method described previously (24). Briefly, spleens from five to eight mice were perfused with RPMI 1640 medium containing 10% heat-inactivated FCS, collagenase (1 mg/ml, type III; Worthington Biochemical, Lorne Laboratories), DNase I (325 KU/ml; Sigma-Aldrich), 0.1 M EDTA, pH 7.2, and 100 U/ml polymyxin B (Sigma-Aldrich). Spleens were digested for 30 min at 37°C, and the homogenate was passed through a 0.40-μm cell sieve. Cells were resuspended in Nycodenz (1.077 g/ml; Invitrogen Life Technologies), layered on Nycodenz, and spun in a Sorval RT7plus centrifuge at 2000 × g for 20 min. DCs were further enriched from the low-density fraction using anti-CD11c microbeads (Miltenyi Biotec). Purity was checked on a FACSCalibur (BD Biosciences) using CellQuest software (BD Biosciences); 98% purity was routinely attained.
DCs were irradiated (2500 rad) and plated in RPMI 1640-DC medium at the indicated numbers per well in 96-well V-bottom plates. CD8+ responder T cells were enriched from the lymph nodes of OTI or LCMV TCR transgenic mice by negative selection; responder populations contained 93–98% CD8+ T cells. A total of 1–2 × 104 responder cells was added per well. In assays using OTI responder cells, the synthetic peptide SIINFEKL was added at 0.1 μM. When LCMV TCR transgenic cells were used as responders, rIL-2 (a gift from J. Tite, GlaxoSmithKline, Stevenage, U.K.) was added at 100 U/ml in the absence of any peptide (25). After 3 days of culture, plates were pulsed for 16 h with [3H]thymidine (1 μCi/well; Amersham Biosciences). [3H]Thymidine incorporation was analyzed following cell harvesting by liquid scintillation counting using a Trilux 1450 Microbeta counter (PerkinElmer Wallac).
For immunocytochemical analysis of the location of DCs within the spleen, alcohol-fixed 5- to 8-μm cryosections of spleen from uninfected and LCMV-infected C57BL/6 mice were double-labeled for DCs and marginal zone (MZ) metallophilic macrophages using biotinylated anti-mouse CD11c (clone HL3; BD Pharmingen) and CR-Fc (a gift from L. Martinez-Pomares, Oxford University, Oxford, U.K.) (26), respectively. CR-Fc is a chimeric molecule consisting of the cysteine-rich region of the mouse macrophage mannose receptor (which binds specifically to MZ metallophilic macrophages in the spleen) and the human IgG1 Fc chain. Bound Abs were detected with HRP-avidin-biotin complex (Vector Laboratories) and alkaline phosphatase-conjugated anti-human IgG (Jackson ImmunoResearch Laboratories). Reactions were developed using the Vector Red kit (pink) for alkaline phosphatase and Vector SG kit (gray) for HRP. Sections were counterstained with methyl green (Vector Laboratories).
For TUNEL staining, cryosections were fixed with 2% paraformaldehyde and triple-labeled for DCs, MZ metallophilic macrophages, and apoptotic cells. DCs and MZ macrophages were identified using the primary detection reagents described above and visualized with streptavidin-AlexaFluor-568 conjugate (Molecular Probes) (red) and anti-human IgG-AlexaFluor-633 (Molecular Probes) (blue), respectively. Apoptotic nuclei were identified using a fluorescence-based TUNEL in situ cell death detection kit (Roche Applied Science) (green).
IFN-α was detected in alcohol-fixed sections with rat anti-mouse IFN-α (HyCult Biotechnology, Cambridge Bioscience), followed by a goat anti-rat secondary Ab labeled with AlexaFluor-488 (Molecular Probes) (green). Costaining to allow identification of DCs was conducted using biotinylated anti-mouse CD11c and streptavidin AlexaFluor-568 (red), as described above.
All experiments included appropriate isotype controls (BD Pharmingen) to ensure specificity of labeling. The results were recorded using software for the Leica TCS NT confocal system (Leica Microsystems).
Flow cytometric analysis and sorting
Cells were stained directly after isolation with Abs against surface markers (purchased from BD Pharmingen) by incubating 2–5 × 105 cells with Ab in PBS, 2% FCS, and 0.1% NaN3. For analysis of ex vivo cytokine production, DCs were first incubated for 4 h in RPMI 1640 medium containing 10% heat-inactivated FCS, 50 μM 2-ME, 100 U/ml penicillin, 100 μg/ml streptomycin, 100 U/ml polymyxin B (RPMI 1640-DC), plus 5 μl/ml GolgiPlug (BD Pharmingen). They were then surface-stained, fixed, and permeabilized; then intracellular cytokine staining was conducted using reagents from BD Pharmingen. Staining was analyzed on a FACSCalibur using CellQuest software. In some experiments, labeled cells were sorted on a MoFlo flow cytometer (DakoCytomation).
Analysis of DC apoptosis ex vivo
Purified CD11c+ DC were placed in culture for 4 h in RPMI 1640-DC medium, and then washed and surface stained, as described above. FITC-Annexin V (Boehringer Mannheim) was added for 15 min on ice in buffer containing 1 M HEPES (Invitrogen Life Technologies), pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, and 1.8 mM CaCl2. Cells were washed thoroughly and fixed, and staining was analyzed as described above.
Analysis of in vivo DC proliferation
Mice were injected i.p. with 1 mg of BrdU (Sigma-Aldrich), and thereafter supplied with drinking water containing BrdU at 0.8 mg/ml. At the indicated time points post-BrdU administration, mice were sacrificed and DCs were isolated from the spleen as described above. BrdU staining was then performed, as described previously (27).
Analysis of IFN-α mRNA expression
B220+ and B220− MoFlo-sorted sDC were resuspended in TRIzol (Invitrogen Life Technologies), and RNA purification was performed following the manufacturer’s instructions. Genomic DNA contamination was removed by treatment with 50 U of RNase-free DNase (Promega) for 30 min at 37°C. RNA was purified by phenol/chloroform extraction and precipitation in acidified sodium acetate, and ethanol. cDNA synthesis was performed using the first strand cDNA synthesis kit for RT-PCR (Roche). A total of 2 μl of RT product was then used for PCR with primers specific for universal IFN-α (as described in Ref.28) and β-actin (5′-TGACGGGGTCACCCACACTGTGCCCATCTA-3′ (sense) and 5′-TAGAAGCATTGCGGTGGAGCATGGAGGG-3′ (antisense)) (MWG Biotec) using TaqDNA polymerase (Promega). PCR products were separated and analyzed on a 2% agarose gel.
DCs were incubated at 2 × 106 per well in a 96-well flat-bottom plate in a volume of 200 μl/well for 18 h at 37°C in 5% CO2. Culture supernatants were then harvested and assayed for IFN-αβ biological activity by measuring their ability to confer resistance to encephalomyelocarditis virus infection upon L929 cells, using a method based on that previously described (29). The IFN activity of the supernatants was quantitated by comparison with a rIFN-α (HyCult Biotechnology) standard curve. Each unit, as expressed in the text, represents 4 international IFN units.
LCMV infection of mice is associated with rapid maturation of spleen DCs to become potent APCs capable of activating naive LCMV-specific CD8+ T cells
In C57BL/6 mice infected i.p. with LCMV Armstrong, virus replicates in systemic sites including the spleen, where infectious virus is detectable from day 1 pi, and reaches peak titers on days 3–4 pi. A strong virus-specific CD8+ T cell response is expanded during the first week pi, and mediates virus clearance within 10–14 days. We addressed the functional activation of sDCs at early time points during this infection. As an important role of DCs is to act as APCs that prime naive T cells, we first investigated the ex vivo APC functions of sDC isolated from LCMV-infected mice 1 and 3 days pi.
Initially, we analyzed the ability of these DCs to stimulate the proliferation of naive OVA-specific CD8+ T cells (derived from OTI TCR transgenic mice), in response to exogenously added cognate peptide. As shown in Fig. 1 A, sDC from uninfected mice stimulated naive CD8+ T cell proliferation in response to exogenously added peptide in a dose-dependent fashion. Notably, sDC isolated from LCMV-infected mice as early as 1 day pi stimulated peptide-specific CD8+ T cell proliferation much more potently than sDC derived from uninfected mice.
To gain insight into the kinetics with which sDC acquire and present LCMV Ags in vivo, we then assessed the ability of sDC from LCMV-infected mice to stimulate the proliferation of naive LCMV-specific CD8+ T cells from LCMV TCR transgenic mice ex vivo. No exogenous peptide was added in these assays; the only Ag present was thus in vivo derived. As shown in Fig. 1,B, naive LCMV TCR transgenic CD8+ T cells only proliferated strongly when incubated with sDC from LCMV-infected mice. sDC from LCMV-infected mice did not stimulate the proliferation of naive OVA-specific CD8+ T cells in the absence of exogenously added peptide (Fig. 1 A), demonstrating the Ag specificity of this response. Spleen DC isolated from mice 3 days post-LCMV infection stimulated LCMV-specific CD8+ T cell proliferation more potently than sDC isolated on day 1 pi, suggesting that they were more activated and/or presented higher levels of the GP33–41 peptide on their surface. Nonetheless, DCs isolated from the spleen of mice just 1 day after infection with LCMV were able to stimulate the proliferation of naive LCMV-specific CD8+ T cells ex vivo, suggesting that these cells rapidly acquire the capacity to start priming LCMV-specific CD8+ T cells in vivo.
DCs migrate into the T cell areas of the spleen during the first 3 days following infection of mice with LCMV
We went on to investigate whether the maturation of sDCs to acquire potent Ag presentation properties following LCMV infection of mice was associated with changes in their location within the spleen. Immunohistochemical staining showed that in uninfected C57BL/6 mice (Fig. 2,A), CD11c+ DCs in the spleen were mainly located in the MZ and in the T cell areas (periarteriolar lymphoid sheath) around the central arteriole, with small numbers scattered in the B cell areas, as previously described (30). Following infection of mice with LCMV, we observed dramatic alterations in the MZ structure, and migration of MZ resident DCs into the T cell areas of the spleen (Fig. 2, B and C). Migration of CD11c+ cells toward the T cell areas in the white pulp was apparent from as early as 24 h after LCMV infection. By day 3 pi, there were striking activation-associated changes in the splenic architecture. MZ metallophilic macrophages were largely absent, and the T cell areas of the follicles were enlarged and contained a high density of CD11c+ cells (Fig. 2 C).
Alterations in the subset composition of DCs within the spleen during acute LCMV infection
We next addressed whether there were changes in the total number and/or subset composition of sDCs during early LCMV infection. Fig. 3,A shows the kinetics of LCMV replication within the spleen, and superimposed upon this, the increase in the total number of splenocytes after infection, which is due in large part to the dramatic expansion in the number of virus-specific CD8+ T cells (13). Notably, however, the total number of cells in the spleen remained constant over the first 3 days pi. We found that not only the spleen size, but also the total number of CD11c+ cells in the spleen did not change over this time frame: it remained close to 4.4 ± 0.25 × 106, which corresponded to ∼2% of total spleen cells (Fig. 3 B).
Previous studies have reported that in the spleen of normal C57BL/6 mice, CD8+, CD4+, and B220+ DC subpopulations are present at a ratio of 1:2:1, respectively (9). In line with this, we found a ratio of 1:2.4:1.6 in uninfected C57BL/6 mice. Interestingly, we observed that this ratio was radically altered as LCMV infection progressed, changing to 1:2.9:3.1 after only 24 h of infection and more dramatically to 1:2.5:7.3 by day 3 pi. These alterations were caused by a combination of a drop in the number of CD11chigh CD8+ and CD4+ DCs in the spleen, with the number of cells of each of these subsets declining ∼2-fold by day 3 pi, and a simultaneous increase in the number of double-negative and CD11clowB220+ sDCs, with the latter comprising ∼40% of all sDCs by day 3 pi (Fig. 3 C).
The decrease in the number of conventional DCs in the spleen following LCMV infection of mice could have been due to migration of DC out of the spleen, phenotypic alteration, and/or cell death. The percentage of CD11c+ cells in the peripheral blood did increase following LCMV infection of mice (from 1.8 ± 0.62 in uninfected C57BL/6 mice to 3.51 ± 1.1 in mice 3 days previously infected with LCMV); however, peripheral blood CD11c+ cell numbers were too low to enable determination of whether this increase was due to up-regulation of CD11c on cell types other than DC (in response to activation), or to an actual increase in the number of DC in the blood (possibly reflecting migration of a proportion of sDCs into the blood).
To investigate whether DC death may contribute to the decrease in the number of conventional DCs in the spleen following LCMV infection, we analyzed the propensity of sDC subsets from uninfected and LCMV-infected mice to undergo programmed cell death (PCD) ex vivo by annexin V staining. When sDC were incubated for a short period ex vivo, conventional CD11chigh sDCs from LCMV-infected mice showed a greater propensity to undergo PCD than their counterparts from uninfected mice (Fig. 4,A). By contrast, there was no detectable difference in the rate at which pDC isolated from uninfected and LCMV-infected mice underwent PCD ex vivo. Bcl-2 expression by conventional CD11chigh sDCs from LCMV-infected mice also decreased during early infection (data not shown). To confirm that DCs were undergoing apoptosis in the spleen in vivo, spleen sections from LCMV-infected mice were analyzed using a combination of TUNEL staining to detect cells undergoing apoptosis (green) and CD11c staining (red) to identify sDC (Fig. 4,B). Costaining for MZ macrophages (blue) was also included. Few apoptotic cells were observed in spleen sections from uninfected mice, but much larger numbers of cells undergoing apoptosis were seen in spleen sections taken from mice both 1 and 3 days post-LCMV infection. Costaining revealed that the majority of these cells expressed high levels of CD11c (Fig. 4 B), suggesting that most of the cells undergoing apoptosis in the spleen in the first 3 days after LCMV infection were conventional sDC.
We also assessed the in vivo turnover of sDC subsets during LCMV infection, by carrying out BrdU incorporation experiments. In uninfected mice, the B220+ sDC subset turns over more slowly than the conventional sDC subsets (31). We found that after a 3-day labeling period, ∼29% of B220+ sDC in uninfected mice were BrdU positive; however, this was increased to ∼55% during acute LCMV infection (Fig. 4,C), most likely reflecting enhanced generation and/or recruitment of newly produced B220+ DCs into the spleen. By contrast, we did not observe an increase in the proportion of recently divided CD4+ or CD8+ DCs in the spleen after LCMV infection (Fig. 4 C). Together, these results suggest that enhanced production and/or recruitment of newly generated B220+ DCs into the spleen contribute to the increase observed in the number of sDC of this subset over the first 3 days after infection of mice with LCMV, while the decrease in CD4+ and CD8+ sDCs occurring at this time may be due at least in part to their acquiring a heightened tendency to undergo apoptotic death, which is not balanced by increased replenishment of these sDC subsets.
Phenotypic activation and cytokine production profile of sDC subsets following LCMV infection of mice
Having shown (Fig. 1) that LCMV infection of mice is associated with rapid maturation of sDCs to become potent APCs capable of activation of naive LCMV-specific CD8+ T cells, we went on to dissect the phenotypic and functional changes that occur in individual DC subsets. Within 1 day of infection, we observed up-regulation of multiple surface activation markers on the CD11chigh CD4+ and CD8+ sDC subsets (Fig. 5). This phenotypic activation of the conventional sDC subsets was sustained over the first 3 days of infection. Plasmacytoid sDC also showed activation-associated phenotypic changes within 1 day pi, but unlike the conventional sDC subsets, were not induced to express high levels of MHC and costimulatory molecules (Fig. 5).
We then examined IL-12 and TNF-α production by sDC subsets from LCMV-infected mice by ex vivo intracellular cytokine staining. Secretion of IL-12 and TNF-α was induced within 24 h of infection (Fig. 6). The kinetics of production of both cytokines was similar, with a higher frequency of cytokine-secreting cells being present on day 1 than day 3 pi. CD11chighCD8+ DC contained the highest frequency of cells producing each cytokine (Fig. 6). In both this and the CD11chighCD4+ DC subset, the number of TNF-α-producing cells was almost twice the number of IL-12-producing cells, while the ratio of TNF-α:IL-12-producing cells in the CD11clowB220+ DC subset was higher. Cytokine production by the CD11clow DC subset was also more transient than that by the CD11chigh subsets, with an elevated frequency of cytokine-producing CD11clow cells being observed on day 1, but not day 3 pi.
We also addressed the involvement of sDC subsets in production of type 1 IFN following LCMV infection of mice. RT-PCR analysis of IFN-α mRNA expression in total CD11c+ sDCs from uninfected and LCMV-infected mice showed that IFN-α mRNA production was up-regulated within 24 h of LCMV infection, and declined to scarcely detectable levels by day 3 pi (Fig. 7,A). Similar analysis on sDC subpopulations separated on the basis of B220 expression indicated that IFN-α mRNA up-regulation was restricted to the pDC subset (Fig. 7,B). Consistent with this, costaining of spleen sections from LCMV-infected mice with Abs against IFN-α and CD11c suggested that the majority of cells producing IFN-α in the spleen on day 1 pi expressed low levels of CD11c (Fig. 7,C). Furthermore, the B220+, but not the B220− subset of CD11c+ sDC was also found to produce type 1 IFN ex vivo (Fig. 7 D). All in all, these results suggest that after LCMV infection of mice, there is a rapid and transient induction of type 1 IFN production by pDCs in the spleen. Other cell types may also contribute to the type 1 IFN response in acute LCMV infection, but the conventional sDC subsets do not play a major role in IFN-α production.
Analysis of the role played by type 1 IFNs in sDC activation and functional maturation in LCMV-infected mice
We previously showed that type 1 IFNs are capable of activating conventional DCs both in vivo and in vitro (18). To investigate the role of type 1 IFNs in mediating the rapid activation of conventional sDCs we had observed during early LCMV infection of mice, we infected IFN-R KO mice and control mice of the same genetic background (129) with LCMV and compared sDC activation and maturation in these animals. In intact 129 mice, there was phenotypic activation of CD11chigh sDC subsets within 24 h of infection, which increased further by day 3 pi. By contrast, in type 1 IFN-R KO mice, the phenotypic activation of CD11chigh sDC observed on day 1 pi was much more modest, and there was relatively little further activation by day 3 pi (Fig. 8,A). Importantly, however, DCs extracted from the spleen of type 1 IFN-R KO mice 1 day after LCMV infection had acquired the capacity to stimulate strong proliferation of naive OVA-specific CD8+ T cells in response to cognate peptide, albeit somewhat less efficiently than DCs from similarly infected 129 mice (Fig. 8 B). Hence, despite their reduced phenotypic activation, these cells had undergone evident functional maturation. Thus, at least in the context of LCMV infection, functional maturation of sDCs can occur by type 1 IFN-independent pathways.
We also addressed the involvement of type 1 IFNs in mediating the reduction we had observed in the number of conventional DCs in the spleen during early LCMV infection. As in C57BL/6 mice, there was a marked drop in the total number of CD11chigh CD4+ and CD8+ sDCs following LCMV infection in 129 mice (Fig. 9,A). Strikingly, however, the dramatic reduction in the number of CD8+ DC in the spleen that occurred in intact 129 mice was not seen in type 1 IFN-R KO mice, and CD4+ sDC numbers were reduced more slowly and to a lesser extent than in 129 mice (Fig. 9,A). Moreover, CD11chighCD8+ DCs from the spleen of LCMV-infected type 1 IFN-R KO mice did not undergo PCD at an enhanced rate ex vivo, and the increase in the ex vivo PCD rate of CD11chighCD4+ DCs from the spleen of day 3 LCMV-infected type 1 IFN-R KO mice was smaller than that seen in similarly infected 129 mice (Fig. 9,B). In line with this, while many apoptotic CD11c+ cells were apparent in the spleen of LCMV-infected 129 mice on day 3 pi, few apoptotic cells were seen in spleen sections from LCMV-infected IFN-R KO mice (Fig. 9 C). These results reveal an important role for type 1 IFNs in mediating the reduction in CD11chigh CD8+ and CD4+ sDC numbers associated with acute LCMV infection.
LCMV Armstrong infection of mice stimulates the induction of a strongly type 1-biased immune response, with dramatic expansion of virus-specific CD8+ CTL, which play a key role in mediating viral clearance, occurring during the first week pi. In this study, we characterized the phenotypic and functional changes in sDC subpopulations that take place during the early stages of this infection. We showed that the initial rounds of LCMV replication in the spleen are accompanied by a rapid functional maturation of all sDC subsets to acquire potent cytokine production and/or Ag presentation properties. We also showed that by day 3 pi, the subset composition of the sDC population is markedly altered, with a reduction in the number of DCs of the conventional CD11chigh subsets, and an increase in the proportion of cells with a pDC phenotype. Furthermore, we addressed the role played by type 1 IFNs in mediating these changes, showing that although they do contribute to the activation of conventional sDC subsets in mice acutely infected with LCMV, and also to their subsequent loss from the spleen, sDC maturation to acquire potent CD8+ T cell stimulatory properties can be triggered during this infection by type 1 IFN-independent mechanisms.
Phenotypic and functional activation of all sDC subsets was evident within 24 h of infection of mice with LCMV, and was accompanied by migration of MZ resident DCs into the T cell areas of the spleen. Activation-associated phenotypic changes in the pDC subset included an increase in MHC and costimulatory molecule expression from their constitutively very low levels to levels similar to those expressed by the conventional DC subsets in uninfected mice. Similar phenotypic activation of pDC has also been found to occur in response to murine CMV (MCMV) infection of mice (32). In their resting state, pDCs are poor stimulators of T cell responses (7, 8), but they have been reported to acquire an enhanced T cell stimulatory capacity following maturation (9, 32, 33). It is thus possible that this DC subset may contribute to expansion of the T cell response in LCMV-infected mice, although as discussed below, the CD11chighCD8+ DC subset most likely played the major role in CD8+ T cell priming in this infection. pDCs have been shown to be a major source of type 1 IFN production in a number of murine virus infections (8, 10, 32); however, previous studies have suggested that pDCs are not the major source of serum type 1 IFN in LCMV-infected mice (10). We found that type 1 IFN production is up-regulated in pDCs in the spleen within 24 h of infection, with mRNA levels subsequently declining by day 3 pi. Previous studies (10) may have failed to detect type 1 IFN production from pDCs from LCMV-infected mice because analysis was conducted 2 days after infection, missing the peak of type 1 IFN secretion by these cells. Although the major source of the type 1 IFN in the serum of LCMV-infected mice, which reaches peak titers 2–3 days pi, remains unclear (10), our observations suggest that pDCs may make an important contribution to local type 1 IFN production in the spleen at the time when the immune response starts to be initiated there.
Conventional sDCs also up-regulated expression of MHC, adhesion, and costimulatory molecules to high levels and began to produce proinflammatory cytokines within 24 h pi, consistent with their playing an important role in priming of the LCMV-specific T cell response. Previous studies have shown that the Armstrong strain of LCMV replicates predominantly in cell types other than DCs within the spleen, with tropism being determined at the level of receptor binding (34, 35), implying that the virus-specific CD8+ T cell response in this infection must be predominantly induced by cross-priming. This would suggest that the CD11chigh CD8+ DC subset, which has been shown to be the principal mediator of cross-priming (5), plays a key role in initiation of the virus-specific CD8+ T cell response in LCMV-infected mice. Consistent with this hypothesis, we have found that in radiation bone marrow chimeras created using bone marrow from rel B knockout mice, which have a deficit in conventional DCs of the CD4+, but not CD8+ subset (36), there is efficient priming of virus-specific CD8+ T cells following LCMV infection (our unpublished data).
It was notable that the number of conventional CD11chigh DCs in the spleen began to drop within 24 h of infection of mice with LCMV, and declined further by day 3 pi (Fig. 3 C). This could have been due to migration of these DC subsets out of the spleen, phenotypic alteration, and/or cell death. We obtained some evidence consistent with the hypothesis that there may be migration of DCs from tissues into the blood during early LCMV infection. In addition, we observed that DCs of the subsets that declined in numbers in the spleen in vivo showed an enhanced tendency to undergo PCD ex vivo, and that there was no increase in the proportion of recently divided CD4+ or CD8+ DCs in the spleen after LCMV infection. This suggests that another mechanism that likely makes an important contribution to the drop in conventional CD11chigh sDC numbers following LCMV infection of mice is an increase in their death rate that is not balanced by enhanced replenishment of these DC subsets. A recent study suggested that in mice infected with Listeria monocytogenes, APCs are rapidly eliminated by the developing CTL response as a negative feedback mechanism to limit the duration of in vivo Ag presentation (37). However, CTL lysis probably did not make a major contribution to the DC decline that we observed in LCMV-infected mice, as we found that LCMV infection of perforin-deficient mice was also associated with a decrease in the number of conventional DCs in the spleen (data not shown). Our results are thus more consistent with the hypothesis that infection-associated signals trigger terminal maturation of conventional sDC subsets not only to undergo a burst of effector activity (possibly associated with some migration from the spleen), but also to acquire an enhanced susceptibility to apoptotic death.
The mechanism(s) by which DCs are activated in LCMV-infected mice is of interest to understand. DC activation may be triggered by direct interaction with viral components and/or indirectly by factors released from host cells. Innate cytokines whose production is rapidly up-regulated in LCMV-infected mice include type 1 IFNs, TNF-α, and a low level of IL-12. Type 1 IFNs are capable of activating conventional DCs both in vitro and in vivo (17, 18) and have potent adjuvant properties (19). We thus addressed the role played by type 1 IFNs in the activation and functional maturation of conventional sDCs in LCMV-infected mice. Comparison of the changes in the phenotype of conventional CD11chigh sDCs during LCMV infection of intact and type 1 IFN-R KO mice showed that type 1 IFNs make an important contribution to the phenotypic activation of these sDC subsets during early LCMV infection, although some phenotypic activation still occurred via type 1 IFN-independent pathways. Importantly, however, sDCs from LCMV-infected type 1 IFN-R KO mice had undergone maturation to acquire the capacity to stimulate efficient priming of naive Ag-specific CD8+ T cells ex vivo (albeit somewhat less competently than sDCs from intact LCMV-infected mice). Consistent with this, previous studies have shown that type 1 IFN-R KO mice are able to mount a virus-specific CD8+ T cell response following infection with LCMV (16). Likewise, limited phenotypic activation of sDC together with (reduced) functional maturation has been observed to occur following MCMV infection of type 1 IFN-R KO mice (32). The mechanism(s) by which DCs are activated in LCMV-infected type 1 IFN-R KO mice is not clear, but may include type 1 IFN-independent actions of viral components on DCs, DC activation via other cytokines, or T cell-induced DC maturation.
Interestingly, the dramatic decrease in the number of conventional DCs in the spleen that was observed following LCMV infection of both C57BL/6 and 129 mice was strikingly dependent on type 1 IFNs. Our results suggest that type 1 IFNs synergize with other DC-activating signals to sensitize conventional sDCs to undergo apoptosis in LCMV-infected mice. It is notable that in MCMV infection, which does not elicit as high a level of type 1 IFN production as LCMV infection (10), activation and functional maturation of sDC subsets take place, but not a subsequent drop in the number of conventional CD11chigh sDCs (32). This again suggests that activation/maturation does not per se lead to a loss of conventional sDCs, but does so only in synergy with type 1 IFN-mediated signaling.
In summary, our results demonstrate the rapid kinetics with which DCs become activated to produce cytokines and acquire the capacity to induce priming of naive LCMV-specific T cells in LCMV-infected mice, but also suggest that optimal conditions for T cell priming are retained for only a short period in vivo, with cytokine production by all sDC subsets being swiftly down-regulated and the number of CD11chigh CD8+ and CD4+ DCs in the spleen quickly starting to decline. Furthermore, we have shown that type 1 IFN production contributes to the phenotypic activation of conventional sDCs in LCMV-infected mice, and to the loss of these sDC. However, our results also demonstrate that there are type 1 IFN-independent pathways for the functional maturation of conventional sDCs, allowing for T cell priming in infections in which type 1 IFN production and/or activity are impaired.
We thank Andrew Worth for cell sorting; Miranda Ashton for excellent technical assistance; Steven Archibald for photographic assistance; Luisa Martinez-Pomares for the gift of CR-Fc; Sandra Diebold and Caetano Reis e Sousa for provision of OTI cells; John Tite for provision of IL-2; and Simon Wong, Agnes Le Bon, and David Tough for critically reviewing the manuscript.
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.
This work was supported by core funding from the Edward Jenner Institute for Vaccine Research. This core funding is provided by the Medical Research Council, Biotechnology and Biological Sciences Research Council, Department of Health, and GlaxoSmithKline. This is Publication 81 from the Edward Jenner Institute for Vaccine Research.
Abbreviations used in this paper: DC, dendritic cell; IFN-R KO, type I IFN receptor knockout; LCMV, lymphocytic choriomeningitis virus; MCMV, murine CMV; MZ, marginal zone; PCD, programmed cell death; pDC, plasmacytoid DC; pi, postinfection; sDC, spleen DC.