Published online before print January 8, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(IFN-)Laboratory of Cellular Therapy, Department of Hematology and Oncology, Children’s University Hospital, Friedrich-Alexander University Erlangen-Nuremberg, Erlangen, Germany
1 Correspondence: Children’s University Hospital, Friedrich-Alexander University Erlangen-Nuremberg, Loschgestr. 15, D-91054, Erlangen, Germany. E-mail: wolfgang.holter{at}uk-erlangen.de
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
generates a more complex pattern of mixed DC phenotypes in response to TLR stimulation. DCs activated by the TLR ligand R-848 in the presence of IFN- and PGE2 produced high levels of IL-12p70 and IL-23, started migration toward CCL19 within only 10 h, and still continued to secrete IL-12p70 without further restimulation following the migration step. The accelerated onset of migration was a result of PGE2 and was associated with reduced plastic adherence and lower amounts of activated CD29. In contrast, IFN- by itself enhanced cell adhesion and strongly hindered CCR7-mediated migration in the absence of PGE2. This suggests a new role for IFN- in the direct regulation of DC migration through enhanced cell adhesion, perhaps to support the development of T cell effector functions at peripheral sites. Together, our data are relevant to the development of DC vaccines, as they demonstrate the existence of dual-functional DCs, which as a result of the simultaneous effects of PGE2 and IFN-, can migrate rapidly toward lymph node chemokines and carry with them a wave of primary cytokines.
Key Words: R-848 • migration • CCR7
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Among such factors individually known to strongly influence DC functions are IFN- and PGE2. IFN- is released in large amounts from activated NK cells as part of the initial, nonspecific immune reaction and supports Th1 priming in lymph nodes [3
]. Regarding DCs, IFN- enhances the release of IL-12p70, in particular, in response to TLR-mediated DC activation [2
, 4
, 5
]. Other IFN- effects on DCs, such as the reduction of IDO expression, have also been reported [6
, 7
].
PGE2 is released by activated monocytes and macrophages at the micromolar and by in vitro-derived DCs at the nanomolar level [8 9 10 ]. At the sites of inflammation, PGE2 has been found at concentrations between 0.2 nM and 1.69 µM [11 12 13 ]. PGE2 is reported to inhibit DC IL-12p70 secretion and seems to be critically necessary for DC migration. It increases the expression of CCR7 and the fraction of migratory DCs [14 15 16 ], but it has also been reported to reduce the migration of DCs through extracellular matrix (ECM) by influencing the balance of metalloproteinases and their inhibitors secreted from DCs [17 ]. PGE2 further inhibits DC IL-12p70 secretion at the site of inflammation [14 , 16 , 18 ], as well as after their migration to the lymph node in response to secondary CD40 ligation [19 ].
Based on these effects of PGE2 on DC cytokine secretion and migration, a well-supported model proposes that maturing DCs acquire a migratory phenotype to induce a T cell response in draining lymph nodes or a so-called proinflammatory phenotype to condition the microenvironment at the peripheral site of inflammation [16
, 20
]. In such a scenario, however, effector T cells infiltrating inflammatory sites would not be supported by DC-derived IL-12p70 as soon as PGE2 has been locally released. Such PGE2 containing inflammatory sites further would probably contain only few mature DCs at all, which are capable of supporting the local expansion of lymphocytes, as matured cells would rapidly have migrated away. On the other hand, these migratory-type DCs would not be able to release significant levels of IL-12p70 upon interaction with lymphocytes in the lymph node and thus, would not be equipped to induce naïve Th1 and CTL cell differentiation, which in vitro, has been shown to be IL-12-dependent [21
, 22
]. Although PGE2-conditioned DCs provided as a cancer vaccine can indeed lead to the expansion of IFN--producing T cells in vivo in cancer patients [23
], given the strictly postulated dependency of DC migration on PGE2, the question remains whether substantial amounts of DC-derived IL-12p70 can be produced within lymph nodes.
We asked how such a scenario would be influenced by the presence of IFN-, possibly provided by NK cells or T cells, which one could imagine to be contained within a distal inflammatory cell infiltrate as well. Therefore, we studied the combined effects of IFN- and PGE2 on the phenotype, cytokine secretion, and migration of monocyte-derived DCs, maturing upon TLR activation. Based on our previous work [2
], R-848 (also called S-28463 and resiquimod) was chosen as a TLR ligand, activating intracellular TLR8 in monocyte-derived DCs. Our experiments showed that TLR8 activation in the presence of IFN- and PGE2 resulted in substantial migration with early onset and allowed secretion of large amounts of IL-12p70 from the migrating DC fraction. Beyond the existence of such a dual-functional DC phenotype, our experiments suggest that the combined action of IFN- and PGE2 shapes a quite complex pattern of mixed DC phenotypes, which might allow DCs to cover changing demands during an ongoing immune response.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(Bender MedSystems, Burlingame, CA, USA), different concentrations of PGE2 (aqueous dilutions freshly prepared from DMSO stocks, Sigma Chemical Co.), or a cytokine cocktail described previously [25
], containing 1000 U/ml recombinant human TNF-
(Bender MedSystems), 1 µM PGE2, 10 ng/ml IL-1β (R&D Systems, Minneapolis, MN, USA), and 1000 U/ml IL-6 (Strathmann Biotec). Where indicated, we added 50 µM forskolin (Sigma Chemical Co.) or 50 µM cAMP analog Sp-5,6-dichloro-1-β-D-ribofuranosylbenzimidazole-3',5'-monophosphorothioate (DCI-cBIMPS; Biolog Life Science Institute, Bremen, Germany) as an alternative to PGE2. Immunophenotypic analysis and functional assays were performed with DCs after the indicated stimulation intervals. To include easily detachable and adherent cells for analysis, cells were routinely harvested by adding 5 mM, pH 7.4, EDTA to the culture medium and incubation for 30 min at 37°C. In some experiments, easily detachable and adherent cells were sequentially harvested for separate analysis. For determination of viable cell numbers, the cells were counted after addition of 5 mM EDTA and propidium iodide (PI) by flow cytometric analysis using counting beads (Invitrogen). Viable cell numbers were determined immediately prior to their further use in functional assays. Secondary IL-12 production was assayed by stimulating 50,000 DCs with 150,000 irradiated CD40 ligand (CD40L) expressing murine fibroblasts (kindly provided by Richard A. Kroczek, Robert-Koch Institute, Berlin, Germany) in the presence of 1000 U/ml IFN- in 500 µl DC culture medium. Light microscopy photographs from the cells were taken directly in the culture dishes using a Nikon Eclipse TS100 inverse microscope (Nikon Corp., Japan) and a 5-Megapixel Nikon Digital Sight DS-5M charged-coupled device camera.
Flow cytometric analysis
For immunophenotypic analysis of the DCs, the following mAb were used: anti-CD1c-PE (clone AD5-8E7; Miltenyi Biotec); anti-CD14-allophycocyanin (clone RMO52) and anti-CD80-PE (clone MAB104; Beckman Coulter, Fullerton, CA, USA); anti-CD33-FITC (clone WM54; Dako, Denmark); anti-CD38-allophycocyanin (clone HIT2; BioLegend, San Diego, CA, USA); anti-CD115-PE (clone 12-3A3-1B10, rat, eBioscience, San Diego, CA, USA); anti-CD1a-FITC (clone HI149), anti-CD83-allophycocyanin (clone HB15e), anti-CD86-FITC (clone 2331), unlabeled anti-CCR7 (clone 2H4), biotinylated rat anti-mouse IgM (clone R6-60.2), streptavidin-PE (all from BD Biosciences, San Jose, CA, USA); and isotype controls: murine IgM (clone G155-228) and IgG2a-allophycocyanin (BD Biosciences); FITC-, PE-, and allophycocyanin-labeled IgG1 (clone DAK-GO1; Dako); and rat IgG1-PE (Invitrogen). For intracellular staining, the cells were incubated with 1% formalin (Carl Roth GmbH, Germany) for 10 min at 4°C and washed once with PBS containing 0.1% saponin (Sigma Chemical Co.) and 50 mM D-glucose (termed "permeabilizing solution") and once with permeabilizing solution supplemented with 10% human serum type AB (HS) (Lonza Walkersville, Walkersville, MD, USA). For surface staining of activated β1 integrin (CD29), the cells were washed once and incubated directly in the culture dishes for 20 min at 4°C with unlabeled anti-CD29 (clone 12G10, AbD Serotec, Raleigh, NC, USA). Staining with secondary polyclonal goat anti-mouse F(ab')2-PE (Dako) was done in the presence of 5 mM EDTA for 30 min at 4°C for simultaneous detachment of the cells. Analysis was performed on a FACSCalibur flow cytometer (BD Biosciences), and data analysis was performed with CellQuest software (BD Biosciences).
Cytokine measurements
Based on standard sandwich ELISA methodology, commercially available pairs of mAb were used to quantify human IL-6, IL-10, and IL-12p70 (all purchased from BD Biosciences); IL-23 (BioSource, Worcester, MA, USA); and TNF- (Diaclone, Stamford, CT, USA). IL-23 protein was also quantified by using an unlabeled capture antibody against the IL-23 subunit p19 (R&D Systems) and a biotinylated detection antibody against IL-12p40 (BD Biosciences) with a detection limit of 200 pg/ml.
Migration assay
DCs were detached from the culture plates by adding 5 mM EDTA, pH 7.4, to the culture medium and incubation for 30 min at 37°C. After washing with PBS 30,000–100,000, cells were resuspended in 100 µl IMDM medium (Invitrogen), supplemented with 5% HS. To avoid reattachment, the resuspended cells were immediately seeded onto migration chamber inserts in 24-well plates (8.0 µm pore size; 0.15x106 pores/cm2, Greiner Bio-One, Germany). The lower compartment contained 600 µl of the same prewarmed medium supplemented with 150 ng/ml CCL19 or 150 ng/ml CXCL12 (PeproTech, Rocky Hill, NC, USA). After 90 min of incubation, 5 mM EDTA was added to the lower compartment, and after a further 30-min of incubation at 37°C, the migrated cells could be detached completely from the bottom side of the porous migration chamber insert, a step that was found to be important to compensate for plastic adherence of TLR-differentiated DCs. After adding PI, the migrated, viable cells were counted by flow cytometry.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and PGE2 on the induction of surface markers in DCs. CD83 and the costimulatory molecules CD80 and CD86 were strongly expressed under all conditions, whereas the chemokine receptor CCR7 was differentially expressed. In accordance with published data, we found that CCR7 expression was heterogeneous on TLR8-activated DCs and could be strongly increased by PGE2 [14
, 16
], whereas IFN- partially reduced CCR7 up-regulation (Fig. 1)
.
 View larger version (27K): [in a new window] |
Figure 1. Immunophenotype of DCs after stimulation with R-848 in the absence or presence of IFN- and PGE2. Immature DCs were cultured in serum-free AIM-V medium and stimulated as indicated on Day 5 of culture. The histograms display the expression of the indicated markers after 2 days of stimulation by R-848 in the absence or presence of IFN- and 1 µM PGE2 (filled histograms). Open histograms with thin lines represent the appropriate isotype and fluorochrome-label controls included in the staining of the cells cultured with PGE2. The histograms show data of a representative experiment; similar data were obtained with three different donors.
|
up-regulated the expression of CD38, similar to what has been reported previously for human monocytes and DCs [30
, 31
]. This IFN--induced expression of CD38 was not significantly changed by the presence of PGE2 (Fig. 1)
. Together, the data confirm that maturing TLR8-activated DCs homogeneously express CD80, CD86, and CD83, a phenotype largely unaffected by additional IFN- or PGE2. In contrast to these maturation and costimulation markers, however, the soluble factors differentially influenced the expression of CCR7 and CD38, molecules previously implicated in cell migration. The expression of the M-CSF receptor CD115 was not induced under any condition, and also, CD1a was not regulated by PGE2 and IFN-.
IFN- and PGE2 enhance the production of IL-12p70 and IL-23 by TLR-activated DCs
Following activation via TLRs, DCs "condition" their surroundings by secretion of a range of chemokines and cytokines. Among these DC-derived cytokines, IL-12p70 and IL-23 are important for a cellular response, as they support the activation of Th1, Tc1, and NK cells [32
]. The amount of released IL-12p70 strongly varies following stimulation of different TLRs but in general, is amplified in the presence of IFN- [2
, 4
, 5
]. Figure 2A
illustrates this amplification for R-848, which although inducing substantial levels of IL-12p70 by its own, leads to a high level of IL-12p70 secretion only in the presence of IFN-.
 View larger version (28K): [in a new window] |
Figure 2. (A) Amplification of the primary production of IL-12p70 and IL-23 depends on PGE2 concentration. Immature DCs were stimulated with R-848 in the absence or presence of IFN- and different concentrations of PGE2 as indicated. IL-12p70 and IL-23 were assayed in the supernatants 48 h later by ELISA. The diagrams show the results obtained with four different donors represented by individual symbols. Immature DCs (untreated) and DCs treated with IFN- alone produced no detectable IL-12p70 or IL-23 (data not shown). Similarly, the amounts of IL-23 produced by R-848 stimulation alone were below the detection limit of the assay (<0.2 ng/ml). (B) PGE2 differently regulates the primary release of different cytokines. Immature DCs were stimulated on Day 5 for 48 h with R-848 plus IFN- in the absence or presence of 10–7 molar PGE2 as indicated. The diagram shows the results obtained with four different donors represented by individual symbols. Where indicated (n.d.=not detectable), IL-10 and IL-23 were below the detection limit of the assays (<20 pg/ml). (C) PGE2 effect on the cytokine release from CD1c-positive myeloid blood DCs. CD1c-positive, CD33-positive DCs were stimulated immediately after their isolation from peripheral blood with R-848 plus IFN- in the absence or presence of 10–7 molar PGE2 as indicated. The diagram shows the results obtained with three different donors represented by individual symbols (detection limit, 20 pg/ml). IL-10 levels of Donor B were 30 pg/ml in the absence and 80 pg/ml in the presence of PGE2. (D) PGE2 effect on primary IL-12p70 release induced by different TLR ligands. Immature DCs were stimulated on Day 5 for 48 h with the indicated stimuli in the absence or presence of 10–7 molar PGE2. The diagram shows the levels of IL-12p70 as percentage values relative to the respective cytokine levels in the absence of PGE2, which was set to 100% (mean±SD; n=4). The 100% levels of IL-12p70 production ranged from 1.74 to 5.20 ng/ml for poly(I:C)/IFN-, 0.14–1.53 ng/ml for LPS/IFN-, and 4.24–20.84 ng/ml for R-848/IFN-. (E) PGE2 effect on primary cytokine release induced by R-848 in combination with different TLR ligands. Immature DCs were stimulated on Day 5 for 48 h with R-848 plus poly(I:C) or LPS in the absence or presence of different PGE2 concentrations. The diagrams show the results obtained with three different donors represented by individual symbols. (F) Effect of a cAMP analog and forskolin on the primary induction of IL-12p70 release by R-848. Immature DCs were stimulated on Day 5 for 48 h with the indicated stimuli in the presence of different cAMP sources. The diagram shows the levels of IL-12p70 as percentage values relative to the respective cytokine levels in the absence of any cAMP source, which was set to 100% (mean±SD; n=4). The 100% levels of IL-12p70 production ranged from 0.28 to 3.70 ng/ml for R-848 and 4.24–20.84 ng/ml for R-848/IFN-. (G) PGE2 effect on the primary induction of IL-12p70 release by R-848 in different culture media. Immature DCs were grown in the indicated culture media and stimulated on Day 5 for 48 h with R-848 plus IFN- in the absence or presence of 10–7 molar PGE2. The diagram shows the results obtained with four different donors represented by individual symbols.
|
, PGE2 has been reported to impair the secretion of primary IL-12p70 [14
, 16
, 18
]. Although PGE2 is synthesized by DCs, only at low levels [8
9
10
], it has been found at inflammatory sites in concentrations between 0.2 nM and 1.69 µM, secreted by monocytes, macrophages, and fibroblasts [11
12
13
]. For these reasons, we were interested how different doses of PGE2 would affect the cytokine secretion of TLR-activated DCs. Unexpectedly, our experiments revealed that intermediate levels of PGE2 resulted in an approximate threefold amplification of IL-12p70 and an approximate tenfold enhancement of IL-23 secretion in the presence of IFN- (Fig. 2A)
. Such an effect of PGE2 has been reported previously only for IL-23 but not for IL-12p70 [33
, 34
]. Figure 2B
shows that the presence of PGE2 also strongly amplified the release of IL-6 but did not result in IL-10 induction or in marked TNF- modulation. To investigate whether our findings could be extended to freshly isolated and ex vivo-stimulated DCs, we studied the effect of PGE2 on the cytokine secretion of CD1c-positive myeloid blood DCs (Fig. 2C)
, which are relatively weak producers of IL-12p70, and there are contrary findings on the IL-12p70 release upon stimulation with R-848 [35
, 36
]. Following stimulation with R-848 plus IFN-, we observed in our system no release of IL-12p70 or IL-23, independent of PGE2. However, substantial amounts of TNF-, IL-6, and IL-10 were secreted, of which TNF-, in accordance with Son et al. [37
], was modestly reduced by 0.1 µM PGE2 (Fig. 2C)
. Next, we investigated the effect of PGE2 on monocyte-derived DCs activated by other TLR ligands, alone or in combinations. Under most of these conditions, high levels of IL-12p70 were induced and with the exception of R-848 plus poly(I:C) stimulation, again enhanced rather than reduced by PGE2 (Fig. 2D and 2E) . As the amplification of IL-12p70 release by PGE2 was unexpected, we tried to raise the level of intracellular cAMP by other agents such as the adenylate cyclase activator forskolin and the cAMP analog Sp-5,6-DCI-cBIMPS, which does not discriminate between the cAMP receptors exchange protein activated by cAMP and protein kinase A. Although both agents were used at high concentrations previously reported to strongly inhibit IL-12p70 production [16 , 34 , 38 ], they were not inhibitory in our culture system (Fig. 2F) . Finally, we asked to what extent different culture media could account for the observed discrepancy to the published data. Indeed, these experiments revealed a partial, albeit not complete, inhibitory action of PGE2 on DCs generated and stimulated in the presence of human or bovine serum but not on DCs cultured in parallel in the serum-free medium AIM-V (Fig. 2G) . Taken together, our data demonstrate that PGE2 differentially regulates the cytokine profile of TLR-stimulated DCs and that also IL-12p70 can be amplified by PGE2, at least under certain conditions.
Influence of IFN- and PGE2 on the plastic adherence of DCs and on the activation of β1 integrin (CD29)
Immature DCs are characterized by the formation of podosomes and an adhesive and low-speed migratory behavior. Following DC activation via TLRs, we transiently observed excessive plastic adherence and a highly pronounced, spindle-shaped morphology (not shown). This strikingly adherent stage gradually resolved during maturation for the majority of cells, which is in agreement with the described dissolution of podosomes under similar conditions [39
]. However, as a large fraction of DCs still remained adherent after 2 days of maturation with R-848, we asked how PGE2 and IFN- would affect DC adherence and if the degree of plastic adherence correlates with the migratory capacity of the cell population. Figure 3A
demonstrates that in the presence of PGE2, the spindle-shaped morphology of adherent DCs was dramatically resolved, whereas IFN- alone led to an even enhanced adherence of essentially all cells. The PGE2 effect was also evident with DCs matured in the presence of IFN-, although the adherence was still far more pronounced in these cultures than in cultures supplied with PGE2 alone. Further experiments revealed that although the general degree of adherence strongly varied among different cell donors, the pattern described above—reduced cell adherence by PGE2 and enhanced adherence by IFN-—was evident with each individual donor.
 View larger version (126K): [in a new window] |
Figure 3. PGE2 and IFN- strongly influence the adhesive properties of DCs. (A) The influence of IFN- and different concentrations of PGE2 on the morphology of cultured adherent and nonadherent DCs. The photographs were taken after 48 h of stimulation (x100 original magnification). (B) The histograms show the expression of indicated markers after 2 days of stimulation with R-848 in the presence of IFN- in the detachable fraction of cells (open histograms with thin lines) or the adherent fraction of cells (open histograms with bold lines). Filled histograms represent isotype controls. The expression of CD83 additionally was validated intracellularly ("intra CD83"). (C) The morphology of the DCs after 2 days of stimulation with R-848 and IFN- (x400 original magnification). The photographs were taken before EDTA treatment from unseparated cells and after removal of the suspendable fraction from the adherent cells following 0, 1.5, and 8 h of incubation with 5 mM EDTA in culture medium at 37°C. The photographs demonstrate the transformation of adherent cells already during detachment into cells with dendritic projections.
|
, both cell populations displayed important criteria typical for classical DCs, such as phenotype and morphology.
In view of these strong visible effects of added PGE2 and of IFN- on cell morphology and plastic adherence, we asked whether the changes were associated with an altered activation level of the integrin system, which is involved in the regulation of cell contact during adhesion and migration. Among integrins, the activated form of β1 integrin (CD29) has been found in podosomes of immature DCs and can be assayed specifically by a well-characterized mAb (clone 12G10) [40
]. Using this antibody, we found that R-848 stimulation by itself but in particular, the adherence-reducing effect of PGE2 was associated with a reduction of activated β1 integrin expression (Fig. 4
).
 View larger version (19K): [in a new window] |
Figure 4. Surface expression of activated β1 integrin on differentially activated DCs. The indicated stimuli were added to immature DCs on Day 5 of culture, and the expression of activated β1 integrin in unseparated total cells was assayed on Day 7. The bars display the relative mean fluorescence intensity (MFI) obtained in five independent experiments using three different donors, as calculated, compared with the MFI obtained with immature DCs within each experiment (mean±SD; R-848 plus PGE2, n=4; otherwise, n=5).
|
and PGE2 on the extent and kinetics of DC migration, which can be released by activated lymphocytes in large amounts, could similarly regulate DC migration.
Following TLR8 activation with R-848 alone, 
60% of the DCs expressed CCR7 (Fig. 1
, and data not shown), and about two-thirds of these CCR7-positive DCs migrated in response to CCL19 (Fig. 5A
). In the presence of PGE2, the fraction of CCR7-positive migratory cells increased (Figs. 1
and 5A)
. This situation of large numbers of TLR8-activated, migratory DCs changed as soon as IFN- was present in the system. IFN- reduced CCR7 expression only partially, but it lowered the fraction of migratory DCs to a small proportion in the absence of PGE2 and to intermediate levels when PGE2 was simultaneously present. Similar data were obtained when the DCs were stimulated with PGN instead of R-848 (not shown).
 View larger version (15K): [in a new window] |
Figure 5. (A) Kinetics of the acquisition of migratory capacity. R-848 ± IFN- ± PGE2 was added to immature DCs on Day 5 of culture, and migratory capacity of the total cell population toward CCL19 was assayed after 8, 10, 16, 24, and 48 h of stimulation. The diagram shows the percentage of migrated DCs without subtracting the respective spontaneous migration values (mean±SD, using four different donors, n=4, except: R-848+IFN-, time-points 16 and 48 h, n=8; R-848+IFN-+PGE2 at time-points 10, 16, and 24 h, n=8; 48 h, n=12, respectively). Spontaneous migration in the absence of chemokine after 48 h of stimulation was 3.37 ± 1.48% for R-848, 21.42 ± 3.63% for R-848 + PGE2, 2.31 ± 1.61% for R-848 + IFN-, and 6.60 ± 3.40% for R-848 + IFN- + PGE2. (B) Kinetics of primary production of IL-12p70. Immature DCs were stimulated by R-848 + IFN- + PGE2 on Day 5 of culture. The diagram shows the relative amounts of IL-12p70 produced after 8, 12, 14, 16, and 24 h of stimulation related to the amounts assayed in the supernatants after 24 h of stimulation of the corresponding donor (mean±SD; n=4). For better comparison, the percentages of migrated cells already depicted in figure 5A
are also included in this figure. (C) Kinetics of induction of CD80 and CD86 expression. Maturation of immature DCs was induced by R-848 + IFN- + PGE2 on Day 5 of culture. After the indicated intervals, the cells were detached, and expression of CD80 and CD86 on total cells was assayed by flow cytometry (displayed as MFI).
|
Capacity of migratory DCs for primary production of IL-12p70
The production of several cytokines in activated DCs has been reported to occur within 6–12 h after triggering [34
, 42
], thus paralleling the acquisition of migratory capacity seen in our experiments (Fig. 5A)
. To directly compare both kinetics, we determined the production of IL-12p70 after R-848 activation of DCs (Fig. 5B)
. As expected, both functions reached their maximum within 16–24 h, which translated into an in vivo situation, could principally allow the deposition of substantial amounts of primary IL-12p70 into lymph nodes.
To directly address the question whether migrating DCs would carry on with cytokine production, despite having migrated away from a putative TLR ligand under the influence of CCL19, DCs stimulated appropriately for 10 h were washed, subjected to a 1.5-h lasting migration assay, and recultered following recovery from the bottom chamber of the migration chamber. As shown in Table 1
, these cells still carried on to release large amounts of IL-12p70 during the following 24 h. Asking whether these migrating DCs would also acquire T cell stimulatory capacity, we quantified CD80 and CD86 expression and again found an approximate tenfold increase in the expression of T cell-activating ligands within 16–24 h, as estimated by MFI measurements (Fig. 5C)
. These results suggest that DCs matured under the influence of a TLR ligand in the presence of PGE2 and IFN- would rapidly acquire the capacity for migration toward lymph nodes and simultaneously changing to a T cell-stimulatory phenotype and releasing IL-12p70, a lymphokine of prime importance for Th1 and CTL development.
|
View this table: [in a new window] |
Table 1. Migratory and Nonmigratory DCs Secrete Primary IL-12p70
|
 View larger version (15K): [in a new window] |
Figure 6. (A) Capacity of migratory and nonmigratory R-848-DCs for secondary production of IL-12p70. Maturation of immature DCs was induced by R-848 on Day 5 of culture. After 24 or 48 h of stimulation, the cells were completely detached by adding 5 mM EDTA and separated into migratory and nonmigratory cells by migration toward CCL19. Both fractions were detached by EDTA again, washed, and restimulated with a CD40L-expressing cell line in the presence of IFN-. IL-12p70 was assayed in the supernatants 48 h later by sandwich ELISA. (B) Influence of PGE2 on the capacity for secondary IL-12p70 production. Maturation of immature DCs was induced by R-848 ± PGE2 on Day 5 of culture. After 10 or 24 h of stimulation, the migratory fraction of the cells was obtained by migration toward CCL19 and was restimulated with a CD40L-expressing cell line in the presence of IFN-. IL-12p70 was assayed in the supernatants 48 h later by sandwich ELISA.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and PGE2, two out of many possible soluble factors from such a scenario, our present study already demonstrates an impressive plasticity of DC function in response to such signals.
Summarizing our results, we propose four prototypic, functional phenotypes of monocyte-derived DCs, which develop depending on the presence or absence of IFN- and PGE2 during maturation. These prototypes are extremes, of course, and functional intermediates are likely to occur. For a better illustration, we photographed cultures representing the four prototypes and characterize them in short in Figure 7
:
 View larger version (53K): [in a new window] |
Figure 7. Plasticity of DC phenotype. This figure depicts prototypic cultures illustrating the morphological and functional characteristics of DC phenotypes resulting from modulation by IFN- and PGE2. The photographs show DCs stimulated on Day 5 of culture for 48 h with R-848 ± IFN- ± 1 µM PGE2 as indicated.
|
II
As soon as PGE2 is added to the system, primary IL-12p70 production is enhanced, and IL-23 is induced (Fig. 2A)
. Regarding adhesion and migration, these DCs ("PGE2/–") round up rapidly and almost completely acquire migratory capacity with enormously shortened kinetics (Fig. 5A)
. In parallel, the capacity to produce secondary IL-12p70 secretion is nearly completely lost in these cells (Fig. 6B)
.
III
IFN- strongly amplifies IL-12p70 primary production in TLR-activated DCs ("–/IFN-"; Fig. 2A
and Lehner et al. [2
]). IFN- up-regulates CD38 (Fig. 1)
but dramatically reduces the size of the migratory DC fraction (Fig. 5A)
. Upon restimulation, these DCs produce large amounts of IL-12 (data not shown).
IV
IFN- reduces the migration of PGE2-costimulated, TLR-matured DCs to intermediate levels ("PGE2/IFN-"). IFN-, however, neither influences the kinetics of acquisition of migratory capacity nor recovers high secondary IL-12p70 secretion. Primary secretion of IL-12p70, IL-6, and particularly, IL-23 is clearly enhanced in the presence of IFN- and PGE2 (Fig. 2A
and 2B)
. This type of DC is "dual-functional," as the cells efficiently migrate within 10–16 h of stimulation while still continuing their primary cytokine secretion (Fig. 5A
and Table 1
).
The existence of DCs with the capacity for secondary cytokine secretion together with migratory capacity has been described previously [45 ]. In an earlier study, we demonstrated this to be apparently a general feature of TLR-activated DCs [2 ]. Others have defined further situations, which allow DCs to develop the dual capacity for primary cytokine secretion and migration in a PGE2- and TLR-independent manner, e.g., by blocking the PI-3K pathway in the context of CD40L-mediated DC activation [38 ].
Regarding cytokine secretion, PGE2 has been reported to impair the secondary and the primary secretion of IL-12p70 [14 , 16 , 18 , 19 ]. Here, we demonstrate that depending on culture conditions, PGE2 unexpectedly can even enhance primary cytokine secretion, including IL-12p70. Our data are the only such reported thus far showing enhanced primary IL-12p70 secretion in the presence of PGE2. However, as we also found a partial inhibitory effect on the release of IL-12p70 in serum-containing culture media (Fig. 2G) , the inhibitory effect of PGE2 appears to be dependent on the presence of further, so far unknown, factors apparently contained in serum. As the culture of DCs in a so-called serum-free-defined medium such as AIM-V as well as in conventional media containing different amounts of human or bovine serum represents artificial, in vitro systems, it can only be speculated how these data can reasonably be extrapolated to the in vivo situation.
Apart from IL-12p70, we found PGE2 to also strongly enhance the production of IL-23, which was found at nanogram quantities only in the presence of IFN- and PGE2. The enhancing effect of PGE2 for IL-23 production is in accordance with the literature [33
, 34
]. As the response of IL-12p70 and IL-23 production appeared to have different PGE2 saturation levels, different amounts of PGE2 will affect the ratio of both cytokines. Our data, furthermore, suggest that PGE2 by up-regulation of IL-23 and IL-6 could generate a milieu abetting the differentiation of Th17 cells, which could mediate protection against extracellular bacteria and fungi but might also be involved in autoimmune diseases [46
]. Again, our observed enhancement of IL-6 secretion by PGE2 stands in contrast with data generated with DCs, cultured with 10% FCS, and activated by CD40L, further illustrating the importance of culture conditions and stimulatory details [38
].
Regarding the inhibitory effect of IFN- on DC migration, our data are in partial accordance with Alder et al. [47
]. Similarly to the report by Mailliard et al [45
], in our hands, this effect was associated with an only partial down-regulation of CCR7 (Fig. 1)
and was strongly counteracted by PGE2 (Fig. 5A)
. Different culture media and maturation by a proinflammatory cytokine cocktail versus a TLR ligand could explain some of the observed differences of IFN- on CCR7 expression. There are some hints as to how IFN- could interfere with the cAMP-signaling pathway triggered by PGE2. For instance, it has been shown that the IFN--stimulated adhesion of monocytes is mediated by the PI-3K [48
], and the PI-3K pathway can induce cAMP-specific phosphodiesterases and thus, can reduce intracellular cAMP levels [49
]. In addition, PI-3K and cAMP further downstream might exert more indirect, opposite effects [50
].
Tight regulation of DC migration during an ongoing immune response could be a central issue. A recent study reported a rapid but only transient DC migration in mice challenged with influenza virus, although the mechanisms underlying this phenomenon were not yet clarified [51
]. Furthermore, in the same model, the DCs were found refractory to migration for several days, which was associated with an impaired immune response to a secondary, alternative stimulus. In our experiments, IFN- strongly hindered DC migration in response to CCL19 and also to CXCL12 (data not shown), suggesting to us that activated lymphocytes as a putative in vivo source of IFN- themselves can directly regulate DC migration also by soluble factors in addition to CD40L stimulation [20
].
Regulation of DC migration by PGE2 has been an important topic over the years. In our study, we already saw substantial migration of DCs also in the absence of PGE2 involving about half of the cell population (Fig. 5A) . Low migration in the absence of PGE2 has been previously ascribed to uncoupling of CCR7 signaling [15 ], illustrating the complexity of this issue. From our experiments, it appears that PGE2 additionally acts by regulating the adhesion molecules of DC, as we observed reduced plastic adherence of DCs in response to PGE2, associated with a decreased expression of activated CD29. This is in agreement with a recent study reporting PGE2 to rapidly dissolve podosomes, which was associated by loss of activated CD29 and acquisition of migratory capacity. CD29 mediates the interaction of immature DCs with ECM components, and a role for this molecule in the retention of DCs in the periphery has been postulated [52 ].
When we took a closer look at the kinetics of CCR7 up-regulation and the acquisition of migratory capacity, we found PGE2 to accelerate both enormously (Fig. 5A
; CCR7 expression data not shown). There is evidence for different pathways regulating CCR7 expression, and accelerated up-regulation of CCR7 expression, for example, has been reported by triggering the triggering receptor expressed on myeloid cell-2/DNAX-activating protein-12 pathway [53
]. As a consequence of accelerated migration, PGE2-DCs would quickly accumulate in lymph nodes. If translatable into an in vivo situation, these kinetic issues might well be of biologic importance by determining the location of primarily and secondarily released cytokines. Published data [34
, 42
] suggest that the secretion of many cytokines, including IL-12p70, seems to roughly parallel the acquisition of migratory capacity as estimated by our data. In cell separation experiments, we could show that migratory DCs (stimulated with R-848 in the presence of IFN- and PGE2) still carry on to produce primary IL-12p70 following a CCL19-mediated migration assay (Table 1)
. Based on this, we postulate that DCs modulated by IFN- and PGE2 might well carry a first wave of IL-12p70 production straight into lymph nodes. Cytokines with a delayed release, as reported for IL-27 [34
], could be transported even more effectively into lymph nodes by such cells.
Although our experiments reveal new and complex effects of IFN- and PGE2 on DC function with likely biological consequences, the situation in vivo is certainly more complex. For instance, depending on the nature of the inflammatory stimulus, the kinetics of IFN- and PGE2 release could be very different from each other in vivo. In a TLR-independent rat model of polysaccharide-induced inflammation, the release of several PGs has been studied and revealed an early and transient peak of PGE2 release [54
]. However, an early presence of NK cells or memory T cells at a peripheral inflammatory site might well result in an immediate release of IFN-, whereas mast cells and macrophages would further shape the kinetics and profile of secreted eicosanoids that all are known to modulate DC function [55
56
57
].
Our in vitro data were generated with monocyte-derived DCs. It has been suggested that monocyte-derived DCs in vivo appear only "on demand" at sites of persistent infection to carry over an essential protective role [58
]. Even if the in vivo differentiation of such monocyte-derived DCs follows similar rules such as those defined in vitro, the final in vivo situation is presumably highly complex and will certainly vary between different infectious and inflammatory conditions. However, our data extend the current models of compatibility and incompatibility of migration and cytokine secretion of monocyte-derived DCs. They suggest that peripheral inflammatory sites, even in the presence of PGE2, could contain significant numbers of mature DCs together with high amounts of primary IL-12p70 and IL-23 if IFN- were released from resident lymphocytes. By integrating different cytokine signals with TLR stimulation, DCs will accommodate to changing demands during inflammation. Knowledge of these mechanisms may also support the further development of monocyte-derived DC vaccines.
Received March 12, 2007; revised October 11, 2007; accepted December 1, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
for T(H)1 priming Nat. Immunol. 5,1260-1265[CrossRef][Medline]-primed IL-12 production by IFN regulatory factor 1 J. Exp. Med. 198,1265-1276 on human monocytes and monocytic cell lines J. Leukoc. Biol. 69,605-612-Type-1 polarized dendritic cells: a novel immunization tool with optimized CTL-inducing activity Cancer Res. 64,5934-5937 dose-dependently inhibits prostaglandin E2-mediated dendritic-cell-migration towards secondary lymphoid organ chemokines Vaccine 24,7087-7094[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  Y. Alvarez, C. Municio, S. Alonso, J. A. S. Roman, M. S. Crespo, and N. Fernandez Cyclooxygenase-2 Induced by Zymosan in Human Monocyte-Derived Dendritic Cells Shows High Stability, and Its Expression Is Enhanced by Atorvastatin J. Pharmacol. Exp. Ther., June 1, 2009; 329(3): 987 - 994. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Boniface, K. S. Bak-Jensen, Y. Li, W. M. Blumenschein, M. J. McGeachy, T. K. McClanahan, B. S. McKenzie, R. A. Kastelein, D. J. Cua, and R. de Waal Malefyt Prostaglandin E2 regulates Th17 cell differentiation and function through cyclic AMP and EP2/EP4 receptor signaling J. Exp. Med., March 16, 2009; 206(3): 535 - 548. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
