Inflammation Conditions Mature Dendritic Cells To Retain the Capacity To Present New Antigens but with Altered Cytokine Secretion Function

Dendritic cells (DCs) are found in peripheral tissues and secondary lymphoid organs displaying an immature phenotype characterized by high endocytic function and low expression of surface MHC class II (MHC II) and T cell costimulatory molecules, required for naive T cell priming. Immature DCs constitutively synthesize MHC II molecules, load them with peptides derived from proteins degraded in endosomes, and transiently display the resulting complexes on the plasma membrane (1). These complexes are short-lived because they are endocytosed and degraded in lysosomes, a process accelerated by ubiquitination of the MHC II β-chain by the membrane-associated ubiquitin ligase, MARCH 1 (2, 3). The constant flow of newly generated MHC II–peptide complexes toward the cell surface via the endocytic route, and from the surface back to lysosomes for degradation, is responsible for the steady-state accumulation of MHC II molecules in endosomal compartments of immature DCs. Immature DCs also express receptors for pathogen-associated compounds (e.g., TLR ligands). Encounter of TLR ligands activates DCs and triggers a maturation program characterized by: 1) transient upregulation of macrocytosis/phagocytosis followed by its downregulation; 2) transient upregulation of MHC II synthesis followed by its near shutdown, a process controlled by changes in the expression of the transcription factor CIITA; 3) downregulation of MARCH 1 expression and MHC II ubiquitination; 4) upregulation of T cell costimulatory molecule expression; and 5) secretion of cytokines and chemokines that facilitate T cell priming (1–5). These features enable mature DCs to efficiently present Ags captured at the time of activation and to activate naive T cells. However, mature DCs are inefficient at presenting most forms of Ags encountered after maturation via MHC I and MHC II (6, 7), an exception being Ags captured with surface receptors (8, 9).

Immature DCs also express receptors for inflammatory molecules (e.g., type I IFN) released by DCs or other cells directly activated by pathogen-associated ligands. These molecules cause the DCs to undergo indirect (or bystander) activation (10). Indirectly activated DCs acquire a mature phenotype, increasing expression of MHC II and T cell costimulatory molecules. However, they do not secrete IL-12 in situations where directly activated DCs do (10); whether this is also true for other cytokines has not been determined. Because inflammatory compounds released at the point of infection can indirectly activate DCs at sites where the pathogen is not yet present, both directly (dir-mDCs) and indirectly activated mature DCs (ind-mDCs) are likely produced during infection (11, 12). The function played by the latter, if any, remains unknown. In principle, the ind-mDCs should not be able to present pathogen Ags because, by definition, these are not available at the site of indirect activation. This fact, and the observation that ind-mDCs do not secrete cytokines, has led to the logical conclusion that ind-mDCs do not participate in immunity against foreign Ags (11, 12).

In this article, we report that ind-mDCs retain their capacity to present subsequently encountered Ags and to prime naive T cells in vivo. They can also retain the capacity to secrete cytokines upon encountering TLR ligands, but their cytokine secretion profile is different from that displayed by dir-mDCs. Our results show that inflammation promotes the generation of conditioned DCs (ind-mDCs) still capable of presenting the same or a different pathogen to the one that triggered the inflammation, but with altered cytokine secretion profiles. This suggests that both directly and indirectly activated DCs contribute to the generation of immune responses against primary infection, and that the magnitude and/or type of immune response elicited by a pathogen can be modulated by pre-existing inflammation.

C57BL/6J, CBA (H-2^k), Ly5.1, TLR9^−/− (13), B6.CH-2^bm1 (bm1), CIITA (H-2^k) (14), OT-I (15), OT-II (16), and MHCII^−/− (17) mice were bred under specific pathogen-free conditions at the Walter and Eliza Hall Institute and at the Bio21 Institute animal facilities. Sex- and age-matched mice (6–12 wk) were used and handled according to the guidelines of the National Health and Medical Research Council of Australia. Experimental procedures were approved by the Animal Ethics Committee, Melbourne Health Research Directorate.

For the generation of mouse chimeras, Ly5.1 recipient mice were gamma-irradiated (2 × 0.55 Gy) and reconstituted with 1 × 10⁶ T and B cell–depleted bone marrow cells of each relevant donor strain at a 1:1 ratio. Chimeras were then treated with 1.1 g/l neomycin (Sigma, St. Louis, MO) in the drinking water for the next 6 wk and were used for subsequent experiments after at least 8 wk from reconstitution. For activation of the DCs, mice were injected with 20 nmol CpG1668 (GeneWorks, Hindmarsh, Australia) 12 h before sacrifice. Chimerism and activation were always tested before or during experiments.

Splenic DCs were isolated as described previously (18). Purity of the DC preparation was at least 75%. For further purification, the cells were fluorescently labeled and sorted using a FACSAria (Becton Dickinson) into the following subsets: CD8⁺ DCs (CD45RA⁻ CD11c⁺ CD8⁺ Sirpα⁻) and Sirpα⁺ DCs (CD45RA⁻ CD11c⁺ CD8⁻ Sirpα⁺), using the Ly5.1 or Ly5.2 markers to distinguish between the wild-type (WT; Ly5.1⁺) and TLR9^−/− (Ly5.2⁺) cells. Purity of the DC subsets was at least 95%. For in vitro experiments, DCs were incubated in DC media (RPMI 1640 medium containing 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10⁻⁴ M 2-ME) at 37°C and 10% CO₂. Where indicated, DC cultures included 25 mg/ml polyinosinic-polycytidylic acid (PIC; Sigma) or 1 μg/ml LPS. In vitro, DCs were generated from bone marrow precursors in culture medium supplemented with Flt3L as previously described (19).

OT-I (K^b-restricted anti-OVA_257–264 CD8 T cells) and OT-II T cells (I-A^b–restricted anti-OVA_323–339 CD4 T cells) were purified from pooled lymph nodes (s.c. and mesenteric) by magnetic depletion of non-CD8 or -CD4 T cells, respectively. T cells were then labeled by incubating them 10 min at 37°C in FCS-free PBS containing 0.1% BSA and 2.5 μM CFSE (Molecular Probes, Eugene, OR) or 5 μM Violet Cell Tracer (Life Technologies, Carlsbad, CA). T cell preparations were routinely >85% pure.

DCs were stained with anti-CD11c (N418), anti-Sirpα (P84), anti-CD45RA (14.8), anti-CD8 (53-6.7), anti-CD86 (PO3.1), anti-MHC II (M5/114), anti-Ly5.1 (A20), anti-Ly5.2 (104), or anti–I-A^k–hen egg lysozyme 48–62 (HEL_48–62; AW3.18) (7)_. For detection of transferred OT-I and OT-II cells, anti-CD4 (GK1.5) or anti-CD8 (53-6.7) and anti–TCR-Va2 (B20.1) were used. Propidium iodide (1 μg/ml) was added to the final cell suspension to stain dead cells. For intracellular staining, cells were incubated in the presence of GolgiPlug (Becton Dickinson) for 4–5 h, surface stained as indicated earlier, labeled with fixable viability dye (eBioscience) to stain dead cells, fixed and permeabilized with Fix/Perm kit (Becton Dickinson), and stained with anti–IFN-γ (XMG1.2) and anti-TNF (MP6-XT22; Becton Dickinson). Samples were acquired on an LSR-Fortessa or LSR-II (Becton Dickinson) and analyzed using FlowJo software (Tree Star, Ashland, OR).

DCs isolated from control or CpG-pretreated CBA mice, or generated in vitro from bone marrow precursors, were treated or not with 3 μg/ml CpG1668 or 1 μg/ml LPS and incubated in U-bottom 96-well plates (CoStar) with titrated amounts of HEL (Sigma) for 16 h. Cells were then stained with anti–I-A^k–HEL_48–62 (7), washed, and analyzed by flow cytometry. Each determination was done in duplicate.

Metabolic labeling, immunoprecipitations, and SDS/PAGE analysis were performed as previously described (20). MHC II and MHC I were immunoprecipitated using a rabbit serum (JV1) (21) and monoclonal anti–MHC I (Y3), respectively.

Cytokine quantification was performed as previously described (22). In brief, DCs were cultured in triplicate for 24 h, and cytokine contents in the supernatants were analyzed using a Bio-Plex cytokine kit according to manufacturer’s instructions (Bio-Rad, Hercules, CA). Cytokines measured with this kit are mouse IL-1α, IL1-β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12(p40), IL-12(p70), IL-13, IL-17, Eotaxin, G-CSF, GM-CSF, IFN-γ, keratinocyte chemoattractant, MCP-1, MIP-1α, MIP-1β, RANTES, and TNF-α. Only cytokines that reached detectable levels were analyzed.

Total RNA from purified cells was obtained using RNeasy Mini Kit (Qiagen, Limburg, the Netherlands). cDNA was then synthesized using SuperScript III Reverse Transcriptase (Life Technologies). SYBR green master mix (Roche) was used in quantitative PCRs on a LightCycler 480 (Roche). Expression of all molecules was normalized to housekeeping gene transferrin receptor 2.

Protocol was adapted from Wilson et al. (23). In brief, DCs were placed on MHC II (N22 mAb)–coated coverslips. They were fixed and permeabilized, then blocked using 10% normal goat serum (Sigma) and labeled with anti–MHC II biotin (N22) and rat anti-Lamp1 (Abcam, Cambridge, U.K.). For detection, SAv-Alexa Fluor 488 (Life Technologies) and goat anti-Rat A594 (Life Technologies) were used. Samples were then stained with 5 μg/ml DAPI and mounted using Prolong Gold Antifade Reagent (Life Technologies). Images were acquired on an LSM700 confocal microscope (Zeiss, Jena, Germany) and analyzed using ImageJ software.

Recipient mice were inoculated i.v. with 1 × 10⁶ Violet Cell Tracer–labeled OT-II cells and 0.1 mg OVA (Worthington, Lakewood, NJ) including or not 0.2 μg LPS (Sigma) as indicated. Sixty hours later, total splenocytes were stained with anti-CD4 and anti-TCRVα2, and resuspended in buffer containing 2.5 × 10⁴ blank calibration particles (Becton Dickinson). The total number of live dividing lymphocytes (propidium iodide⁻, Violet Cell Tracer^lo) was calculated from the number of dividing cells per 5 × 10³ beads.

bm1 spleen cells were gamma-irradiated (1500 rad), washed, incubated with 10 mg/ml OVA in RPMI 1640 medium for 10 min at 37°C, and washed three times with RPMI 1640 medium supplemented with 3% FCS, to prepare cell-associated OVA.

DCs were plated in DC media at 1 × 10⁴ cells/well in U-bottom 96-well plates (Costar, Cambridge, MA). For MHC II Ag presentation, DCs were then pulsed for 1 h at the indicated concentration of soluble OVA (Worthington). Cells were then washed and resuspended with 5 × 10⁴ CFSE or Violet Cell Tracer–labeled OT-II cells in DC media including 10 ng/ml GM-CSF (PeproTech, Rocky Hill, NJ). For the cross-presentation assay, DCs were incubated in the presence of titrated amounts of OVA-coated splenocytes (OCSs) and 5 × 10⁴ CFSE-labeled OT-I cells in DC media including 10 ng/ml GM-CSF (PeproTech). Proliferation of OT-II or OT-I T cells was determined after 60 h of culture as explained before for in vivo proliferation assays. For cytokine staining, cultures were left for 72 h and were incubated in the presence of GolgiPlug (Becton Dickinson) and 1 μg/ml OVA_323–356 peptide (Mimetopes, Melbourne, Australia) for 5 h before intracellular staining of the cells and analysis. Each determination was done in duplicate.

Data were plotted using GraphPad Prism (La Jolla, CA). Mann–Whitney unpaired test with two-tailed p values and 95% confidence intervals and two-way ANOVA test were used for all statistical analyses as indicated.

To compare the functional properties of dir-mDCs and ind-mDCs generated in vivo within the same mice, we produced mixed bone marrow chimeras where lethally irradiated WT recipients were reconstituted with a 1:1 ratio of Ly5.1 WT and Ly5.2 TLR9^−/− bone marrow. The chimeric animals contained equivalent numbers of WT and TLR9^−/− DCs (Supplemental Fig. 1), which expressed similar levels of the classical maturation markers CD40, CD86, and MHC II (Fig. 1A). i.v. injection of the TLR9 ligand CpG induced direct activation of the WT DCs within 12 h, becoming dir-mDCs with high expression of the maturation markers (Fig. 1A). The TLR9^−/− DCs of these same mice also acquired a mature phenotype (ind-mDCs), in this case, because of release of inflammatory mediators produced by cells that responded to CpG (Fig. 1A). However, as previously shown (10), ind-mDCs did not secrete IL-12, whereas dir-mDCs did (Fig. 1B). This was also the case for IL-6, MIP-1α, and RANTES (Fig. 1B). Secretion of 12 additional cytokines and chemokines also showed marked differences between ind-mDCs and dir-mDCs, although these differences did not achieve statistical significance (Fig. 1B).

We have previously shown that normal mice cannot mount anti-OVA CD4 T cell responses if they are inoculated with CpG i.v. at least 9 h before immunization (7). This is because CpG causes direct activation of DCs systemically, and by the time of immunization they have already acquired a mature phenotype incapable of presenting the newly encountered OVA Ag (7). We tested whether this was also the case in mouse bone marrow chimeras reconstituted only with WT (WT:WT) or with both WT and TLR9^−/− (WT:TLR9^−/−) bone marrow (Fig. 2A). The mice were injected or not with CpG and 12 h later they were immunized with OVA plus LPS. MHC II Ag presentation was monitored by the adoptive transfer of Violet Cell Tracer–labeled anti-OVA, I-A^b–restricted OT-II cells. As expected, the OT-II cells proliferated in mice that had not received CpG before immunization, but they did not in WT:WT bone marrow chimeric mice pretreated with CpG (Fig. 2A) because of impaired Ag presentation by the systemically activated dir-mDCs. In contrast, CpG pretreatment did not prevent OT-II proliferation in WT:TLR9^−/− mixed bone marrow chimeras (Fig. 2A). This suggested that the TLR9^−/− ind-mDCs contained in these animals were still able to present OVA.

It should be appreciated that in the WT:TLR9^−/− mixed bone marrow chimeras treated with CpG, only half of the DCs (the ind-mDCs) were potentially capable of Ag presentation, so induction of OT-II proliferation in these animals could not be expected to be as efficient as in nontreated controls in which all of the DCs were fully functional. Nevertheless, it was possible that the systemic inflammation induced by CpG caused additional effects that limited the capacity of the DCs to capture Ag, or to interact with OT-II cells in vivo, so we also measured Ag presentation in vitro. This allowed us to compare the effect of CpG on the two major populations of conventional DCs, distinguished by their expression of CD8 (CD8⁺ DCs) and Sirpα (Sirpα⁺ DCs), respectively, because the latter play a more prominent role in MHC II presentation (24, 25).

We purified WT or TLR9^−/− CD8⁺ and Sirpα⁺ DCs from WT:TLR9^−/− mixed bone marrow chimeras that had been left untreated or injected i.v. with CpG 12 h previously. The four DC populations were incubated in vitro with soluble OVA and CFSE-labeled OT-II T cells, and OT-II proliferation was measured 60 h later (Fig. 2B). Both WT and TLR9^−/− DCs from untreated animals were immature at the time of purification. These DCs underwent spontaneous maturation in vitro during the Ag presentation assay (26) and induced OT-II proliferation similarly (Fig. 2B). As shown previously, the Sirpα⁺ DCs were more efficient than the CD8⁺ DCs (24, 25). The DCs purified from CpG-treated animals were already mature (Fig. 1) and, as expected, the dir-mDCs (WT) had impaired capacity to present newly encountered OVA (Fig. 2B). However, Sirpα⁺ ind-mDCs (TLR9^−/−) maintained their Ag presentation capacity almost intact (Fig. 2B). Because the dominant DC subtype involved in MHC II presentation in vivo is the Sirpα⁺ DCs, the results of these in vitro experiments explain the incomplete but significant maintenance of MHC II presentation in CpG-treated WT:TLR9^−/− chimeric mice, where half the DCs were ind-mDCs.

We also assessed IFN-γ and TNF production by T cells that were activated by dir-mDC and ind-mDC Sirpα⁺ DCs in the Ag presentation assay in vitro. The percentage of proliferating OT-II cells that produced these cytokines was similar in all cases (Fig. 2C). Therefore, the main difference between dir-mDCs and ind-mDCs appeared to be at the level of Ag presentation, which determined how many naive T cells were primed and started to proliferate (Fig. 2B), but the rate of proliferation and differentiation of the primed T cells was similar regardless of which type of mDC caused the priming.

Indirect activation seemed to have a stronger effect on MHC II presentation by CD8⁺ than by Sirpα⁺ DCs (Fig. 2B). A possible explanation for this difference was that inflammatory cytokines caused full activation of CD8⁺ DCs but not Sirpα⁺ DCs. Alternatively, it could be that because CD8⁺ DCs are inherently less capable of MHC II presentation, a small reduction in this capacity appeared to have a bigger overall impact on this population. We investigated this question further by assessing MHC I cross-presentation of cell-associated Ag, an activity restricted to the CD8⁺ DC population (27, 28). We purified WT or TLR9^−/− CD8⁺ DCs from WT:TLR9^−/− mixed bone marrow chimeras that had been left untreated or injected i.v. with CpG 12 h previously. The four DC populations were incubated in vitro with OCSs and CFSE-labeled OT-I cells, and proliferation was measured 60 h later (Fig. 2D). Immature WT and TLR9^−/− DCs showed a similar capacity to cross-present OCSs. As shown previously (6), dir-mDC (WT) obtained from mice pretreated with CpG failed to cross-present (Fig. 2D). However, the ind-mDCs (TLR9^−/−) retained most of their capacity to phagocytose and cross-present cell-associated Ag (Fig. 2D). These results support the conclusion that CD8⁺ DCs suffered only partial loss of their capacity to capture and (cross)present Ags via MHC I and II, although the relative effect on MHC II presentation was more pronounced.

We have previously attributed the impairment in MHC II Ag presentation in mDCs to downregulation of MHC II synthesis (7), but formal demonstration that this is the major mechanism has been lacking. MHC II production is transcriptionally regulated by CIITA, whose expression is driven in DCs by a promoter that is silenced in the mature state (29). We examined MHC II Ag presentation by DCs from CBA (H-2^k) mice that, in addition to the endogenous CIITA, also express CIITA as a transgene controlled by an independent, constitutive promoter (14). These DCs upregulate surface MHC II expression during maturation like non-transgenic (non-Tg) DCs, but maintain MHC II synthesis in their mature state (Fig. 3A, 3B). We injected WT or CIITA Tg mice with CpG, and 12 h later purified their splenic DCs (which at this stage were dir-mDCs). Immature DCs were also purified from noninjected mice. The four groups of DCs were incubated for 12 h in vitro with HEL, and formation of I-A^k molecules loaded with the HEL_48–62 peptide was determined by FACS using the mAb AW3.18 (30). The magnitude of staining increased above background level in both WT and CIITA Tg DCs that were immature at the beginning of the assay, indicative of formation of I-A^k–HEL_48–62 complexes during the incubation in vitro (Fig. 3B, 3C). The WT dir-mDCs did not generate such complexes as expected, but CIITA Tg dir-mDCs did (Fig. 3B, 3C). Similar results were obtained with DCs generated in vitro from bone marrow precursors in the presence of Flt3L, activated with either CpG or LPS (Supplemental Fig. 2). Therefore, the impairment in MHC II presentation in dir-mDCs is primarily caused by shutdown of MHC II production (7, 20, 31) and can be rescued by sustaining MHC II synthesis.

Next, we compared MHC II production in dir-mDCs and ind-mDCs purified from untreated or CpG pretreated mixed bone marrow (WT:TLR9^−/−) chimeras. MHC II protein synthesis by dir-mDC (WT) was nearly shut down, but ind-mDCs (TLR9^−/−) from these same mice synthesized similar amounts to their immature DC counterparts (Fig. 4A). As shown previously, MHC I synthesis was upregulated in mature DCs, and this was the case for both dir-mDCs and ind-mDCs (Fig. 4A).

The differences in MHC II expression between dir-mDCs and ind-mDCs were also evident at the transcription level, examined by quantitative real-time PCR (Fig. 4B), and correlated with changes in CIITA transcription (Fig. 4B). Directly activated mDCs also downregulate transcription of MARCH 1 (32, 33), a ubiquitin ligase that enhances turnover of MHC II–peptide complexes, but this was not the case in indirectly activated ind-mDCs (Fig. 4B). Expression of MARCH 2 was examined as a negative control, and indeed no differences were observed in MARCH 2 transcription between dir-mDCs and ind-mDCs (Fig. 4B). We observed that the transcription of MHC II, CIITA, and MARCH 1 was slightly more downregulated in CD8⁺ ind-mDCs than in their Sirpα⁺ counterparts, explaining the differential capacity of the two DC subsets to present Ag via MHC II encountered after indirect activation (Fig. 2B).

In immature DCs, continuous formation of new MHC II–peptide complexes, followed by transient cell-surface expression, ubiquitination, and delivery to lysosomes for degradation, causes a characteristic accumulation of MHC II in intracellular Lamp⁺ compartments (Fig. 4C). After direct activation of WT DCs by in vivo CpG treatment, MHC II was localized at the cell surface in dir-mDCs, but ind-mDCs displayed intracellular MHC II (Fig. 4C). We conclude that, despite the increased surface expression of MHC II and other characteristic markers of DC maturation, the MHC II Ag presentation pathway of ind-mDCs remains essentially as in immature DCs.

Next, we tested whether ind-mDCs are still able to recognize and respond to newly found pathogen-associated signals. WT:TLR9^−/− mixed bone marrow chimeras were left untreated or were injected with CpG to induce formation of dir-mDCs and ind-mDCs. After 12 h, these mice were immunized with OVA with or without LPS. The immunized mice also received CFSE-labeled OT-II cells to measure T cell proliferation in vivo. We observed little OT-II proliferation in mice that were immunized with OVA alone, but extensive proliferation in mice immunized with OVA plus LPS (Fig. 5A). This was the case both in untreated mice and in mice that had been pretreated with CpG. Because the only DCs capable of Ag presentation in CpG-pretreated mice were the ind-mDCs (Fig. 2), this result indicated that ind-mDCs responded to LPS, which increased their capacity to prime OT-II cells.

To further characterize how indirect activation affected the capacity of DCs to respond to a subsequent encounter of TLR ligands, we purified immature WT and TLR9^−/− DCs from untreated WT:TLR9^−/− mixed bone marrow chimeras, and dir-mDCs and ind-mDCs from chimeras injected 12 h prior with CpG. The four groups of DCs were then incubated in vitro with LPS or PIC. After 24 h, we measured the secretion of cytokines and chemokines. The pattern of cytokine secretion induced by LPS on immature WT (Supplemental Fig. 4A) and TLR9^−/− (Fig. 5B) DCs was comparable. Secretion of cytokines by WT cells that first encountered CpG in vivo and subsequently LPS in vitro could be attributed to a direct effect of the first or the second TLR ligand (Supplemental Fig. 4A). The most striking result was observed with ind-mDCs (TLR9^−/−) stimulated in vitro with LPS, which secreted much higher levels of IL-6, MIP-1α, MIP-1β, and keratinocyte chemoattractant than their immature counterparts (Fig. 5B). IL-1α and MCP-1 were also secreted at statistically significant higher levels by the ind-mDCs, although the magnitude of the difference was not as dramatic (Fig. 5B). In contrast, secretion of RANTES in response to LPS appeared inhibited in ind-mDC, although the difference did not reach statistical significance (Fig. 5B). These results show that although inflammatory stimuli were not sufficient to induce cytokine secretion by ind-mDCs (Fig. 1B), they conditioned these cells to change their pattern of cytokine secretion in response to subsequent encounter of a pathogen-associated ligand such as LPS. Conditioning did not affect the response to all TLR ligands. The pattern of cytokine secretion induced by PIC in WT and immature TLR9^−/− DCs was comparable and different from that induced by LPS (Supplemental Figs. 3 and 4B), but this was little altered by pre-exposure of TLR9^−/− DCs to inflammatory cytokines in vivo (Supplemental Fig. 3).

Pathogens can simultaneously induce two DC maturation pathways in vivo. DCs that directly encounter the pathogen become dir-mDCs; those responding to secondary inflammatory signals become ind-mDCs. It is well established that the function of dir-mDCs is to present the pathogen Ags and to prime naive T cells, but the function of ind-mDCs has been less clear (11, 12). In this article, we showed that ind-mDCs retain their capacity to present Ags encountered after the onset of maturation, both through MHC II and (for CD8⁺ DC) MHC I. We also showed that ind-mDCs can secrete T cell stimulatory cytokines in response to TLR ligands, but their pattern of cytokine secretion was altered by the conditioning effect of the inflammatory signals.

Downregulation of MHC II synthesis and turnover enable dir-mDCs to present for long periods pathogen Ags captured in the mature state, but also prevents presentation of newly encountered Ags (7). Our analysis of DCs from CIITA Tg animals formally proved that downregulation of MHC II transcription and protein synthesis is the main mechanism responsible. Conversely, continued CIITA, MHC II, and MARCH 1 expression enable ind-mDCs to present newly encountered Ags. Similar conclusions were reached by Simmons and colleagues (34) in studies of the effect of inflammatory signals on DCs in vitro, and by Lutz (35) using cultured human DCs. Our results demonstrate for the first time, to our knowledge, that this is also the case in vivo and may explain why, in apparent contradiction to our own results, some reports concluded that mature DCs generated in vivo by TLR ligand inoculation retained their capacity to present new Ags (36). In these studies, the dose of TLR ligand used or the route of inoculation may have led to generation of both dir-mDCs and ind-mDCs, with the latter retaining their capacity for MHC II presentation.

CD8⁺ and Sirpa⁺ DCs are functionally distinct (37). CD8⁺ DCs have a high cross-presenting capacity and are the main producers of IL-12, whereas Sirpa⁺ DCs show better capacity to present Ags in the context of MHC II to CD4 T cells (24, 27). Our results show that this functional dichotomy is further exacerbated by indirect activation, because ind-mDCs CD8⁺ became less capable of Ag presentation via MHC II compared with their immature counterparts, but their capacity for MHC I cross-presentation remained largely intact.

By definition, ind-mDCs have not encountered, and therefore do not present, the pathogen Ag that triggered their differentiation (11, 12), but they may contribute to present that same pathogen if they encounter it subsequently. This implies that during any infection, the pathogen may be presented to naive CD4 T cells by two “waves” of mature DCs: first dir-mDCs and then ind-mDCs that later encounter the pathogen and undergo further differentiation. We term this second wave of fully mature DCs 2mDC (Fig. 6). Importantly, the cytokines secreted by dir-mDCs and 2mDC may be different, and so may be the type of CD4 Th response they elicit. The result may be a more heterogeneous Th response than would be expected if only dir-mDCs participated in Ag presentation (Fig. 6). This conclusion may have implications for vaccine design. For example, a vaccine consisting of an Ag coupled to CpG may induce secretion of a limited number of cytokines by DCs in vitro (e.g., IL-12), but in vivo that same vaccine may elicit other cytokines, which, in turn, may induce the formation of Th cells with unforeseen polarity. Another important implication of our results is that an ongoing immune response against a pathogen may influence the response against a coinfection (12). Likewise, underlying inflammatory processes may affect the type of immune response against an infection (12).

To evaluate the contribution of dir-mDCs, ind-mDCs, and 2mDC to immune responses in the scenarios outlined earlier, it will be important to identify differentially expressed molecules that may enable their discrimination and enumeration. At present, cytokine secretion and expression of Ag presentation genes are the only parameters that enable distinction of dir-mDCs and ind-mDCs, but they are not reliable for accurate quantitation. Definition of suitable surface markers for FACS analysis is now an important goal that will facilitate full characterization of the complex DC network in the steady-state and in different scenarios of infection and/or inflammation.

We thank Drs. Luc Otten and Hans Acha-Orbea (University of Lausanne, Lausanne, Switzerland) for the CIITA mice.

The authors have no financial conflicts of interest.