Experimental studies in monkeys on the basis of ex vivo-generated, reinjected dendritic cells (DCs) allow investigations of primate DC biology in vivo. To study in vitro and in vivo properties of DCs with a reduced capacity to produce IL-12, we adapted findings obtained in vitro with human cells to the rhesus macaque model. Following exposure of immature monocyte-derived monkey DCs to the immunomodulating synthetic polypeptide glatiramer acetate (GA) and to dibutyryl-cAMP (d-cAMP; i.e., a cAMP enhancer that activates DCs but inhibits the induction of Th1 immune responses), the resulting DCs displayed a mature phenotype with enhanced Ag-specific T cell stimulatory function, notably also for memory Th1 cells. Phosphorylation of p38 MAPK was not induced in GA/d-cAMP-activated DCs. Accordingly, these cells secreted significantly less IL-12p40 (p ≤ 0.001) than did cytokine-activated cells. However, upon restimulation with rhesus macaque CD154, GA/d-cAMP-activated DCs produced IL-12p40/IL-23. Additionally, DCs activated by proinflammatory cytokines following protocols for the generation of cells used in clinical studies secreted significantly more IL-23 upon CD154 restimulation than following prior activation. Two days after intradermal injection, GA/d-cAMP-activated fluorescence-labeled DCs were detected in the T cell areas of draining lymph nodes. When similarly injected, GA/d-cAMP as well as cytokine-activated protein-loaded DCs induced comparable Th immune responses characterized by secretion of IFN-γ, TNF, and IL-17, and transiently expanded FOXP3+ regulatory T cells. Reactivation of primate DCs through CD154 considerably influences their immmunostimulatory properties. This may have a substantial impact on the development of innovative vaccine approaches.
Interleukin-12 is a heterodimeric cytokine, with the biologically active form (IL-12p70) consisting of an L chain (IL-12p35) and an H chain (IL-12p40). IL-12p40 (together with a unique p19 subunit) is also part of IL-23, an IL-12-like cytokine, which is predominantly involved in the pathogenesis of autoimmune diseases, such as experimental autoimmune encephalomyelitis (1), collagen-induced arthritis (2), psoriasis (3), and T cell-mediated colitis (4). IL-23 drives the development of Th cells that secrete IL-17 (Th17) (5), a proinflammatory cytokine, and, like IL-23, it is implicated in the pathogenesis of autoimmune disorders (6). IL-12 is mainly produced by dendritic cells (DCs)4 or monocytes and favors the differentiation of Th1 cells. However, some recent data argue against the absolute requirement of IL-12 for the induction of Th1 responses (7).
The feasibility to generate human DCs from peripheral monocytes (8, 9) has greatly facilitated studies on primate DC biology (10). PGE2, TNF, IL-1β, and IL-6 efficiently activate immature DCs to produce IL-12p40 but not IL-12p70 (11, 12). Nevertheless, DCs stimulated with those cytokines strongly activate Th1 T cells in vitro (13), prime naive CD4+ T cells toward the Th1 phenotype in vivo (14), and are currently under investigation as immunotherapeutic tools in a large number of clinical studies (15, 16).
Several protocols have been published for the generation of IL-12-impaired DCs. Glatiramer acetate (GA, copolymer-1, Copaxone) is a synthetic random polypeptide that exerts Th2-promoting effects and has been licensed for use in humans (17). DCs incubated in the presence of GA do not secrete IL-12 in response to LPS, which preferentially primes Th1 responses by itself (18), and drive Th2 immune responses in vitro (19). Likewise, enhancers of intracellular cAMP levels (e.g., cholera toxin, forskolin, and dibutyryl-cAMP (d-cAMP)), induce maturation of DCs, inhibit production of IL-12 (20), and support Th2 priming (21). No data exist on the in vivo capacities of such cells (e.g., whether they have the potential to reverse existing Th1 responses), which would be particularly interesting as a potential treatment of autoimmune diseases (22).
Rhesus macaques provide a useful model to study primate immunology and infectious diseases, for example, SIV infection (23, 24, 25). Monkey DCs can be generated from PBMCs (26, 27), and cytokine-activated rhesus macaque DCs secrete high amounts of IL-12p40 and behave in the same way in vitro and in vivo as do human DCs (27, 28). This sets the stage for studies of DC biology, in particular how different ways of DC preparation may influence the properties of these cells in vivo following reinjection, but also for preclinical vaccine studies (29, 30, 31, 32, 33).
For the present study we adapted protocols for the production of human IL-12-impaired DCs to the rhesus macaque system and used GA in combination with d-cAMP as an additional activation stimulus, since for DCs to induce Ag-specific immune responses, their activation followed by maturation is required (34, 35, 36, 37, 38, 39, 40). We established the generation of IL-12-impaired monkey DCs in vitro and analyzed their potential to activate naive and memory T cells in vitro. We compared in these analyses IL-12-impaired with cytokine-activated, IL-12p40-secreting cells, looked at the effect of CD154 (CD40L) as a subsequent stimulus on preactivated DCs, and investigated whether these cells were able to migrate and to induce protein-specific Th cell responses in vivo.
In vitro, mature DCs also induce regulatory T cells (41), which comprise a heterogeneous group of T cells that can suppress Th effector functions in allergy and autoimmune disease, immunity against pathogens, tumors, and transplants, and are involved in immunologic tolerance (42). Besides analyzing peripheral protein-specific immune responses after the injection of Ag-loaded DCs, we therefore also investigated whether reinjected DCs might affect the numbers of peripheral regulatory T cells in vivo. Herein we show that GA/d-cAMP-activated monocyte-derived monkey DCs lack the production of IL-12; however, upon restimulation with CD154, they secrete IL-12p40/IL-23 just like cytokine-activated DCs. In vivo, both subsets are able to induce Ag-specific Th cells secreting proinflammatory cytokines and to transiently expand regulatory T cells.
Materials and Methods
Healthy adult male rhesus macaques housed at the German Primate Center (Göttingen) and Ab-negative for simian T-lymphotropic virus type 1, simian D-type retrovirus, and SIV were used. All animal care operations were conducted in compliance with the guidelines of the German Primate Center and approved by the appropriate authorities.
Multimerized human and rhesus macaque CD154 and control supernatants
293T cells were grown in DMEM supplemented with 10% FCS and 1% penicillin/streptomycin (Invitrogen). Cells were plated into 75-cm2 cell culture flasks (Greiner Bio-One) at a density of 4 × 106 cells/flask. After 24 h, cells were transfected with 30 μg plasmid DNA of either pEF-human rMegaCD154, pEF-rhesus rMegaCD154 (provided by Apoxis), or pGreenLantern-1 (Invitrogen) applying the calcium/phosphate coprecipitation method. Twelve to 16 h posttransfection the media were changed to serum-free DMEM supplemented with 5% AIM-V (Invitrogen). Supernatants were harvested after 48 h and cell debris was removed by centrifugation (1000 rpm, 10 min). Supernatants were filtered through 45-μm filters (Sarstedt) and 20-fold concentrated in Vivaspin 20 columns (Vivascience) with a molecular mass cutoff of 10,000 kDa. Concentrated supernatants were stored at −20°C.
DCs and T cells
For reinjection studies, monkeys were treated for 9 days with 15 μg (1.5 × 106 U)/kg body weight/day human rG-CSF (filgrastim, Neupogen; Amgen) before blood collection. This treatment increases monocyte yields around 2- to 3-fold compared with baseline levels (M. Eisenblätter, E. Jasny, K. Mätz-Rensing, C. Stahl-Hennig, R. Ignatius, manuscript in preparation). Rhesus macaque monocyte-derived DCs were generated from heparinized peripheral blood as previously described (26, 27). CD14+ monocytes were magnetically separated (Miltenyi Biotec) and cultured at 1.5–2 × 106 cells/3 ml in RPMI 1640 medium, supplemented with 2 mM l-glutamine, 50 μM 2-ME, 10 mM HEPES, penicillin (100 U/ml), streptomycin (100 μg/ml) (all Invitrogen), human rGM-CSF (1000 U/ml, sargramostim, Leukine; Berlex Laboratories), human rIL-4 (100 U/ml, R&D Systems), and 5% human AB serum (PAN Biotech). Human monocytes were magnetically separated from PBMCs obtained from buffy coats (German Red Cross) or healthy donors and cultured at 3 × 106 cells/3 ml under equivalent conditions, except that the medium contained 10% FCS (Biochrom) instead of human serum. At day 6, DCs were stimulated for 24 or 48 h at 1 × 105/well in 96-well round-bottom plates or 5 × 105/well in 24-well plates with either human rIL-6, rTNF, rIL-1β (all 10 ng/ml, R&D Systems), and 10 μM PGE2 (Sigma-Aldrich) (11, 27) or with GA (10 μg/ml, Copaxone, Aventis Pharma) and d-cAMP (1 mM, Sigma-Aldrich). Numbers of viable large DCs were determined by trypan blue exclusion. The shape of the cells was documented by digital photography on a Zeiss Axiovert 40 microscope.
For restimulation experiments with multimerized human or rhesus macaque MegaCD154, DCs were activated by GA/d-cAMP or cytokines for 24 h as described above. Supernatants were then discarded and the cells were extensively washed and resuspended in either a supernatant from cells expressing MegaCD154, from GFP-expressing control cells, medium alone, or LPS at a concentration of 1 μg/ml. In some experiments, a specific inhibitor of p38 MAPK (SB203580, Sigma-Aldrich) was used at 25 μM (43). Human DCs were activated with GA/d-cAMP or cytokines, washed twice with medium, and preincubated with SB203580 or DMSO (1/200) for 30 min at 37°C before addition of medium alone or MegaCD154 supernatant. Levels of IL-12p40 and IL-23 in the supernatants 48 h after restimulation were determined by ELISA.
The phenotype of DCs was characterized by flow cytometry using PE- or FITC-labeled anti-human mAbs with known cross-reactivity to rhesus macaque HLA-DR, CD14, CD16, CD25, CD32, CD80, CD86 (all BD Pharmingen), CD11b, CD83 (both Caltag Laboratories), CD11c (BioSource/Invitrogen), CD40 (Beckman Coulter/Immunotech), CD64 (Serotec), and CCR7 (R&D Systems) or the appropriate isotype controls (BD Pharmingen) (27). Cells were analyzed on a FACSCalibur cytometer with CellQuestPro software (BD Pharmingen). The endocytic functions of DCs were studied by using dextran-FITC (10 μg/ml) and transferrin-FITC (20 μg/ml) (both Molecular Probes/Invitrogen) as described (27).
To enrich rhesus macaque or human T cells, 1 × 108/ml PBMCs were incubated with FITC-conjugated anti-human HLA-DR mAbs (BD Pharmingen), followed by anti-FITC-mAb-conjugated magnetic beads (Miltenyi Biotec). HLA-DR-negative cells were used as responder cells in allogeneic and autologous T cell proliferation assays. Purity of T cells as monitored by staining with PE-conjugated mAbs to CD3, CD14, and CD20 (all BD Pharmingen) was ≥90% in all experiments.
Analysis of p38 MAPK phosphorylation
Immature human DCs (1.5 × 106) were incubated with cytokines or GA/d-cAMP at 37°C for 15, 30, and 60 min; unstimulated cells were used as controls. Cells were washed twice with ice-cold PBS, resuspended in 100 μl cell lysis buffer containing protease and phosphatase inhibitors (Sigma-Aldrich), and incubated for 60 min on ice. Cell debris was removed by centrifugation for 30 min at 14,000 × g and 4°C. Supernatants were harvested and stored at −80°C. Equal amounts of supernatants (10 μl per lane) were mixed with 10 μl of 2× SDS sample buffer (Bio-Rad), boiled for 5 min, subjected to 10% SDS-PAGE, and transferred to Trans-Blot nitrocellulose membranes (Bio-Rad). Membranes were blocked with blocking buffer (5% nonfat dry milk in Tris-buffered saline) for 1 h at room temperature and incubated overnight at 4°C with mAbs to phospho-p38 or total p38 MAPK (Cell Signaling Technology). Bands were visualized with horseradish-peroxidase-conjugated mAbs by an ECL system (Cell Signaling Technology) (43).
T cell assays and intracellular cytokine staining
DCs generated from two rhesus macaques previously immunized with keyhole limpet hemocyanin (KLH) (44) or from a human donor vaccinated with Mycobacterium bovis bacillus Calmette-Guérin (BCG) were activated in the presence or absence of 10 μg/ml endotoxin-free KLH (Calbiochem) or 5 μg/ml purified protein derivative of Mycobacterium tuberculosis (PPD). DCs were added in graded doses to 2 × 105 autologous T cells in flat-bottom 96-well tissue culture plates (Nunc). DCs and T cells alone served as controls, and T cell proliferation was assessed as described below. Supernatants were harvested after 5 days.
PBMCs (105 cells/well) were incubated in triplicates with 100 μg/ml KLH or medium alone in 96-well round-bottom plates. Controls were incubated in the presence of staphylococcal enterotoxin B (SEB; 5 ng/ml; Alexis) or Con A (5 μg/ml; Sigma-Aldrich). Supernatants were harvested after 48 h. After 5 days (3 days for SEB and Con A) of incubation, [3H]thymidine (1 μCi/well, PerkinElmer) was added, and its incorporation was measured on a liquid scintillation counter (Tri-Carb 2000CA, PerkinElmer) 24 h later.
In some experiments 5 × 105 PPD-loaded or unloaded DCs were added to 1.5 × 106 autologous CFSE-stained (0.5 μM) T cells in flat-bottom 48-well tissue culture plates (Nunc) in 1 ml medium, and restimulated with or without PPD (5 μg/ml) at day 6. Cells were incubated for 6 h in polystyrene tubes coated first with goat anti-mouse IgG Ab (H+L, affinity-purified F(ab′)2 fragments, Kierkegaard and Perry Laboratories) followed by costimulatory Abs to CD28 and CD49d (both BD Pharmingen) (45). Brefeldin A was added for the last 4.5 h and extracellular staining with anti-CD3 PerCP and anti-CD4 allophycocyanin (both BD Pharmingen) was performed followed by fixation in 4% paraformaldehyde/PBS and permeabilization (0.5% saponin in PBS plus 5% FCS and sodium azide 10 mM). After washing and resuspension in 0.5% saponin, the cells were incubated for 1 h at room temperature with PE-conjugated Abs against IFN-γ (clone 4S.B3, BD Pharmingen), IL-17 (clone eBio64CAP17), or CD154 (clone 24–31, both eBioscience) for intracellular staining. Cells were washed in 0.5% saponin and fixed in 4% formalin/PBS. Samples were analyzed by flow cytometry after gating on lymphocytes; 10 events were acquired for each sample.
Cytokine/chemokine concentrations were quantified by using sandwich ELISAs for the detection of monkey IL-2, IL-4, IL-10, IL-12p40, IFN-γ, or human IL-12p70 (all U-CyTech Biosciences), CCL3 (Antigenix America), IL-10 (in DC supernatants; Bender MedSystems), IL-17, IL-23 (both eBioscience, NatuTec), or IL-18 (MBL International).
Immature DCs were stimulated with GA/d-cAMP for 24 h and stained with 5-chloromethylfluorescein diacetate (CMFDA; CellTracker Green, Molecular Probes/Invitrogen) (46). Before reinjection, the cells were centrifuged and resuspended in 1 ml of PBS. Donor animals were injected with DCs intradermally at three sites (total volume 1 ml) ∼1–2 cm from palpable inguinal or axillary lymph nodes. Thirty-six hours after injection of CMFDA-labeled cells, all draining lymph nodes were surgically removed, embedded in OCT medium (TissueTek, Invitrogen), snap frozen, and stored at −80°C.
To induce Ag-specific immune responses, immature DCs were treated with pyrogen-free KLH (10 μg/ml) and GA/d-cAMP or cytokines for 24 h before washing and injection. Injections of KLH-loaded DCs were repeated up to three times (Table I). All DC preparations were kept on ice before injection and injected within 5 h after harvesting from cell culture.
|Stimulus .||Animal No. .||Weeks .||.||.||.||.|
|.||.||0 .||10 .||14 .||16 .||20 .|
|GA/d-cAMP||12741||1.2 × 106||1 × 106||1 × 106|
|12745||1 × 106||1.5 × 106||1.2 × 106|
|Cytokines||12743||1.2 × 106||1 × 106||1.2 × 106|
|12744||6 × 105||1 × 106|
|Stimulus .||Animal No. .||Weeks .||.||.||.||.|
|.||.||0 .||10 .||14 .||16 .||20 .|
|GA/d-cAMP||12741||1.2 × 106||1 × 106||1 × 106|
|12745||1 × 106||1.5 × 106||1.2 × 106|
|Cytokines||12743||1.2 × 106||1 × 106||1.2 × 106|
|12744||6 × 105||1 × 106|
Fluorescent microscopy and immunohistochemistry
Frozen lymph nodes were sectioned at 8–10 μm, fixed in 95% alcohol for 30–60 s, mounted with Fluoromount G (SouthernBiotech, obtained through Biozol Diagnostica) and investigated using an Olympus BX60 microscope with fluorescence unit BX-FLA (both from Olympus). For immunohistochemistry, cryostat sections were fixed in acetone for 10 min and incubated with a mAb to fluorescein (1/300, Molecular Probes). Ab binding was visualized by the alkaline phosphatase-anti-alkaline phosphatase method using New Fuchsin as chromogen. The sections were counterstained with hemalaun, mounted, and examined for the presence of fluorescein-positive cells. To characterize the maturation of the injected DCs, double immunolabeling was performed. Cryostat sections were incubated with the mAb to fluorescein as described above. Binding of the Ab was detected with the immunoperoxidase method. After fixation of the sections in 2% paraformaldehyde for 40 min, they were boiled in 0.01 M buffered sodium citrate solution (pH 6.0) for 1 min to unmask Ags. Sections were incubated overnight with mAbs against either CD83 (dilution of 1/100, Immunotech) or CD1a (dilution of 1/20, BD Pharmingen), and binding of mAbs was visualized through the alkaline phosphatase anti-alkaline phosphatase method using fast blue as chromogen. The sections were mounted without counterstaining.
KLH-specific T cell lines
To generate KLH-specific T cell-lines, PBMCs were cultured in 96-well flat-bottom trays (Nunc) in cell-culture medium in the presence of KLH (10 μg/ml). Recombinant IL-2 (50 U/ml, aldesleukin, Proleukin; Chiron) was added at day 5, and T cell lines were analyzed at day 20 (47).
For intracellular cytokine staining, T cells were stimulated with 10 ng/ml PMA and 1 μg/ml ionomycin (Sigma-Aldrich) in 200 μl medium in 96-well round-bottom trays (Nunc). After 1 h of incubation, 10 μg/ml brefeldin A (Sigma-Aldrich) was added to the cultures for the last 5 h. Cells were harvested and stained with mAbs against human CD3 (allophycocyanin-conjugated) and CD8 (FITC-conjugated, both BD Pharmingen) followed by intracellular staining with IFN-γ (clone 4S.B3), IL-4 (clone 8D4–8), IL-10 (clone JES3–9D7), or TNF (clone MAb11, all BD Pharmingen), as described above, and analyzed by flow cytometry.
Data were analyzed using Student’s t test or Mann-Whitney U test.
GA/d-cAMP activated DCs display a mature phenotype and down-regulate their endocytic activity
As in the case of immature human DCs (19), the immunomodulating peptide GA did not activate rhesus macaque DCs (data not shown). We therefore combined GA with a cAMP enhancer (i.e., d-cAMP), which activates DCs but inhibits IL-12 production (20). Similar to data obtained with human cells (20), a d-cAMP concentration of 1 mM efficiently activated monkey DCs (data not shown). Monocyte-derived monkey DCs activated with either the combination of GA and d-cAMP or proinflammatory cytokines differentiated in culture into large, veiled cells (Fig. 1,A), as described for mature DCs (50). GA/d-cAMP and cytokine-activated DCs displayed an identical phenotype regarding expression of surface molecules, including high expression of the homing receptor to secondary lymphoid tissues, CCR7 (51) (Fig. 1,B). Both populations expressed low levels of CD11b, CD16, and CD64, as well as moderate levels of CD11c and CD32 (Fig. 1 C). A similar phenotype was observed in GA/d-cAMP-stimulated human DC cultures (data not shown). Both stimuli led to comparable yields of mature DCs; that is, 60–80% of the cells seeded into the wells could be recovered after 48 h.
The high endocytic activity of immature DCs is down-regulated upon maturation (52, 53). We therefore tested the effects of GA/d-cAMP on endocytic DC functions. Activation of monkey DCs with either GA/d-cAMP or cytokines resulted in a complete loss of uptake of dextran, while receptor-mediated uptake of transferrin was markedly reduced as well (Fig. 1,D). In contrast, when immature cells were incubated in the combined presence of maturation stimulus and dextran or transferrin, endocytosis was not affected by the stimuli (Fig. 1 E), reflecting simultaneous Ag capture and maturation as in previous vaccine studies (28). Thus, rhesus macaque DCs that are mature by a large number of criteria, including phenotype and Ag uptake, can be generated using the combined stimulus of GA and d-cAMP.
Impaired IL-12p40 but sustained CCL3 secretion by GA/d-cAMP-activated DCs
To test whether IL-12 production is inhibited in GA/d-cAMP-stimulated DCs, we quantified IL-12 in cell culture supernatants obtained 48 h after stimulating rhesus macaque DCs with GA/d-cAMP or cytokines. GA/d-cAMP-activated DCs produced significantly less IL-12p40 (p ≤ 0.001) than did cytokine-treated cells (Fig. 2,A), and the same effect was seen in identically stimulated human DC populations (data not shown). In contrast, there was comparable secretion of CCL3 (p = 0.49; Fig. 2 B), while we detected neither IL-10 nor IL-18 in supernatants of DCs stimulated for 48 h with either stimulus (data not shown). Likewise, the addition of GA/d-cAMP to DCs simultaneously treated with proinflammatory cytokines suppressed the production of IL-12p40 significantly (cytokines plus GA/d-cAMP: mean of 351 (range 95–889) pg/ml; cytokines: mean of 3170 (range 1168–6064) pg/ml; p = 0.0003).
The IL-12p40 production of human DCs induced by CD154 or TNF is signaled via the p38 MAPK phosphorylation pathway (43). Because numbers of monkey DCs were insufficient to prepare cell lysates for studies of the p38 MAPK phosphorylation state of GA/d-cAMP-activated DCs, we determined this in cell lysates of monocyte-derived DCs of human origin. Stimulation of immature DCs with cytokines was followed by a transient but sharp increase in phosphorylation of p38 MAPK with a maximum at 15 min, which decreased to baseline levels within another 45 min. In contrast, no increased p38 MAPK phosphorylation was detected after stimulation of immature DCs with GA/d-cAMP (Fig. 2, C and D). Phenotypic activation of control cells following incubation for 48 h was documented for both DC populations (data not shown). Thus, the significantly reduced IL-12 production by GA/d-cAMP-activated DCs corresponded with a lack of p38 MAPK phosphorylation.
Although there is no method available to selectively determine monkey IL-12p70, cytokine-activated monkey DCs produced IL-12 most likely in the form of IL-12p40, since also human cytokine-activated DCs secrete IL-12p40 but not p70 (Ref. (12) and our observations). As IL-12p40 is also part of the recently described cytokine IL-23 and could have been detected in the IL-12p40 ELISA used by us, we further investigated whether cytokine activation of primate DCs leads to production of IL-23. Immature DCs obtained from 10 rhesus macaques were incubated in the presence of the proinflammatory cytokines, supernatants were collected after 48 h, and concentrations of IL-12p40 and IL-23 were determined by ELISA. While the cells produced considerable amounts of IL-12p40, only a small fraction of IL-12p40 had apparently been secreted in the form of IL-23, as we detected only low concentrations of this cytokine in the DC supernatants (Fig. 3). Thus, cytokine-activated primate DCs produce IL-12p40 predominantly in the form of IL-12 while only a minor part is due to the formation of IL-23.
Activation of memory Th cells
The process of DC maturation is generally accompanied by an increased T cell stimulatory capacity (26, 54, 55), and both GA/d-cAMP-activated and cytokine-matured macaque DCs stimulated the proliferation of human T cells in allogeneic MLR assays more efficiently than did immature DCs (data not shown). To determine whether GA/d-cAMP-matured DCs were able to activate Ag-specific memory T cells, DCs were generated from two previously KLH-immunized monkeys, loaded with KLH during activation, and cocultured with autologous T cells. GA/d-cAMP-matured DCs induced specific T cell proliferation to a comparable extent as cytokine-treated DCs (Fig. 4,A). Supernatants from these co-cultures contained substantial amounts of IFN-γ (Fig. 4,B), whereas IL-4 or IL-10 was not detected (data not shown). Likewise, both GA/d-cAMP- and cytokine-activated, PPD-loaded DCs from BCG-vaccinated individuals induced proliferation (Fig. 5,A) and IFN-γ secretion (Fig. 5,B). On the single cell level, autologous CD4+ T lymphocytes proliferated and produced IFN-γ, while no IL-17 was detectable (Fig. 5 C). The production of IFN-γ was accompanied by expression of CD154 by comparable proportions of T cells (mean 13.3 ± 11.3% for cytokine-treated vs 21.3 ± 15.3% for GA/d-cAMP-treated DCs). Together with our previous data, these results indicate that GA/d-cAMP- and cytokine-matured DCs are equally effective at capturing, processing, and presenting protein Ag and subsequently activating Ag-specific memory Th1 cells.
Both GA/d-cAMP-activated DCs and cytokine-stimutaled DCs secrete IL-12p40/IL-23 upon CD40 ligation
Functions of preactivated DCs (e.g., of cells reinjected for therapeutic purposes) might be additionally influenced through subsequent stimuli to which the cells are exposed (e.g., CD154 or proteins of the extracellular matrix). We therefore studied the effects of a subsequent CD40 ligation on the cytokine profile of GA/d-cAMP-activated vs cytokine-stimulated DCs in vitro. In contrast to the lack of substantial IL-12p40 secretion directly after GA/d-cAMP activation (Fig. 2,A), the subsequent incubation of GA/d-cAMP- and cytokine-activated DCs with rhesus macaque CD154 now stimulated the production of comparable amounts of IL-12p40 by both DC populations (Fig. 6,A). To determine whether p38 MAPK plays a role in the restoration of IL-12 production after secondary stimulation, we added a specific inhibitor of the p38 MAPK during the CD154 restimulation (43). This drastically reduced the production of IL-12 by DCs and was independent of their primary stimulation with cytokines or GA/d-cAMP (Fig. 6 B).
Since we did not detect IL-12p70 in supernatants from human DCs set up alongside with the monkey cells in these experiments, we again determined IL-23 and detected considerable amounts of this cytokine at equal concentrations in supernatants of GA/d-cAMP- or cytokine-activated, CD154 restimulated monkey DCs (Fig. 6,C). Notably, cytokine-activated, not restimulated DCs, had produced comparable amounts of IL-12p40 as CD154 restimulated DCs (p = 0.37) but significantly lower levels of IL-23 than restimulated cells (p < 0.01; Fig. 3).
IL-12 production in GA/d-cAMP-prestimulated DCs could not be restored by restimulation with LPS (Fig. 6 D), suggesting that a T cell signal may be required to overcome the deficiency rather than any strong DC-activating signal.
Thus, restimulation with CD154 but not a TLR4 ligand restores the ability of GA/d-cAMP-treated DCs to produce IL-12, and this process involves p38 MAPK. CD154 restimulation also elicits the production of significant amounts of IL-23 by both, primarily cytokine- and GA/d-cAMP-stimulated DCs.
Migration of GA/d-cAMP-activated DCs in vivo
To monitor migration patterns of GA/d-cAMP-matured DCs, we injected CMFDA-labeled DCs from four monkeys intradermally into the donor animals, as this route was superior to s.c. application in a recent study with human monocyte-derived DCs (56). Visualized by immunohistochemistry, cells were distributed throughout the T cell areas of the lymph nodes (Fig. 7,A), most likely entering through the sinuses (Fig. 7,B). In contrast, B cell follicles were free of fluorescein-positive cells. The migrated DCs were mature, as shown by the absence of CD1a (Fig. 7,C), a molecule predominantly expressed on immature DCs and the co-expression of the DC maturation marker CD83 (Fig. 7,D, arrows), and they were located close to resident DCs (Fig. 7, C and D, blue, single-positive cells). A similar distribution of fluorescein-positive cells was seen in the other three animals investigated either by fluorescence microscopy or immunohistochemistry.
Peripheral CD4+ T cell responses induced by DC immunization
Using the same KLH-naive monkeys as for the migration studies, we were able to perform a small pilot study to investigate the potential of GA/d-cAMP-activated DCs to induce CD4+ T cell-mediated immune responses. DCs were loaded with KLH, simultaneously activated with GA/d-cAMP or cytokines, and reinjected into the donor animals. After the first injections only animal 12743 (cytokine-matured DCs) showed KLH-specific immune responses 1 wk after injection; that is, low concentrations of IFN-γ (38.1 pg/ml) were detected in the supernatant of KLH-stimulated PBMCs. After the second and third immunizations, moderate KLH-specific proliferative responses were apparent (recipients of cytokine-activated DCs: stimulation index (SI) ranging from 1.5 to 4.7; recipients of GA/d-cAMP-activated DCs: SI ranging from 1.9 to 2.9). We detected IL-2 in supernatants from KLH-stimulated PBMCs obtained from all animals after the second and, more prominently, after the third immunization (Fig. 8, upper panel). After three immunizations, IFN-γ was detectable in cell culture supernatants from animal 12741 and in those from animal 12743 (Fig. 8, lower panel). IL-4 or IL-10 was not found in the supernatants but detectable in those of SEB/Con A-stimulated control cells (data not shown).
To more accurately analyze the KLH-specific responses, we repeatedly established CD4+ T cell lines 2–6 mo after the immunizations. Given the limited possibility to repeatedly draw large amounts of blood from rhesus monkeys, which is required to generate DCs, we chose a protocol published for the analysis of immune responses induced by KLH-loaded DCs in humans that does not require repeated restimulation with DCs (47). While the percentages of IFN-γ-secreting cells differed at the various time points of establishing the cell lines, the pattern of cytokine secretion was identical in all four animals; that is, we detected considerable percentages of IFN-γ-secreting cells in cell lines from the animals 12741 and 12743, and less IFN-γ producing cells in those from animals 12744 and 12745 (Fig. 9,A), which had yielded lower responses also in the previous assays. In none of the animals, substantial numbers of IL-4 secreting cells, were detectable (Fig. 9 B). In two of the three experiments we looked also for IL-10- and TNF-secreting cells, and we did not detect significant differences between the animals (IL-10: 12743, 3.8 and 1.6%; 12744, 0.8 and 3.9%; 12741, 5.1 and 7.9%; 12745, 4.2 and 1.4%; TNF: 12743, 20.6 and 21.9%; 12744, 10.9 and 25.9%; 12741, 20.6 and 32.7%; 12745, 14.6 and 16.2%). Thus, while GA/d-cAMP-activated primate DCs without restimulation considerably differed from cytokine-activated DCs regarding their IL-12 secretion in vitro, both cell populations were able to induce Th immune responses in vivo that were characterized by high numbers of IFN-γ- and TNF-producing cells and low numbers of IL-4- and IL-10-secreting cells.
Both DC subsets secreted IL-23 upon restimulation by ligation of CD40 (Fig. 6). To investigate whether a possible in vivo production of IL-23 might have given rise to IL-17-secreting Th cells (57), we analyzed the IL-17 concentrations in the supernatants of the CD4+ T cell lines. IL-17 was detectable in supernatants of cell lines from all animals, and there was no clear trend toward higher IL-17 concentrations in supernatants from either group of animals (Table II). Hence, both IL-12p40-secreting and IL-12p40-impaired primate DCs are able to induce comparable Ag-specific Th immune responses in vivo.
|DC Stimulus .||Animal No. .||Cell Line 1 .||Cell Line 2 .||Cell Line 3 .|
|DC Stimulus .||Animal No. .||Cell Line 1 .||Cell Line 2 .||Cell Line 3 .|
PBMCs were cultured in 96-well flat-bottom trays in cell culture medium in the presence of KLH (10 μg/ml), recombinant IL-2 was added at day 5, and supernatants were harvested at day 20.
Transient expansion of regulatory T cells after the injection of cytokine- or GA/d-cAMP-activated DCs
While we did not detect any changes within the subpopulation of FOXP3+ T cells in PBMCs after the first DC injections, a 3.9-fold increase of regulatory T cells was detected at week 4 compared with week 0 after the second injection in animal 12744 (which was injected only twice in total). Similarly, the other animals showed increased numbers of FOXP3+ T cells after the third immunization (animal 12741: 7.2-fold, animal 12745: 3.1-fold, and animal 12743: 2.9-fold increase at week 1 compared with week 0) (Fig. 10). Thus, the DC reinjections transiently expanded peripheral regulatory T cells in the donor animals irrespective of the stimulus used to activate the DCs.
Several protocols have been published for the generation of DCs with significantly impaired IL-12 production. We therefore sought to investigate properties of such cells in an established primate model in vitro and in vivo. Stimulation of immature rhesus DCs with GA and d-cAMP reliably activated the cells, as determined by surface marker expression and down-regulation of Ag uptake. GA-d-cAMP-activated primate DCs lacked IL-12 production to a similar extent as did cells activated by GA and LPS (19) or by d-cAMP alone (20), and this inability to produce IL-12 was also found in cells simultaneously treated with proinflammatory cytokines. The impaired production of IL-12 was accompanied by the lack of phosphorylation of p38 MAPK. In contrast, p38 MAPK was phosphorylated in cytokine-activated DCs, and the cells produced IL-12, most likely in form of IL-12p40, since cytokines alone do not induce production of IL-12p70 (12). In fact, two signals (e.g., CD40 ligation plus IFN-γ or LPS) are required for efficient induction of IL-12p70 secretion in DCs (58). Furthermore, IL-12p70 production is inhibited by PGE2 (59). Cytokine activation of primate DCs led to only marginal production of IL-23. Nevertheless, IL-12p40 alone also possesses immunostimulatory functions (60). GA/d-cAMP-activated DCs were still able to produce other cytokines than IL-12, including the Th1 cell-activating chemokine CCL3 (61).
Both GA/d-cAMP- and cytokine-activated monkey and human DCs were able to stimulate protein-specific memory Th1 cells. While they did not produce IL-18, this property of GA/d-cAMP-activated DCs may result from the maintained production of CCL3—and probably other T cell-activating chemokines/cytokines—as well as from the slight IL-12p40 secretion. Our results are in agreement with recent findings by Schnurr et al. (12) who showed that IL-12 secretion was significantly impaired in adenosine triphosphate-treated human DCs; however, such cells activated memory Th1 cells in vitro, and this could be reduced by IL-12 neutralization. Th1 cells at a later stage of their life-cycle may be activated by fewer amounts of IL-12 or may be less dependent on IL-12 for their activation, as demonstrated in immunized mice, where neutralization of IL-12 did not reduce Ag-specific production of IFN-γ by spleen cells 2 wk after the immunization (62). The results of our recall assays furthermore demonstrate that T cells stimulated by autologous Ag-loaded DCs express CD154 and may thus bypass the p38 MAPK pathway (43). In Th2 immunity, IL-12-overexpressing DCs prevented the development of Th2 pathology in murine allergic lung disease but could not reduce preexisting eosinophilic airway inflammation, which was even enhanced (63). Collectively, the data indicate that IL-12-impaired DCs cannot be used therapeutically to silence or reverse preexisting Th1 immune responses in vivo (e.g., in autoimmune diseases).
Activated DCs (e.g., DCs that have picked up Ag in the periphery) may additionally receive subsequent stimuli that might affect their Ag-presenting properties. Molecules of the extracellular matrix have been shown to influence DC biology (64). Alternatively, the cells could be affected by CD154 expressed on T cells. To mimic this scenario, we investigated the effects of a subsequent CD40 ligation on the cytokine profile of GA/d-cAMP or cytokine preactivated DCs. Surprisingly, GA/d-cAMP-activated DCs now secreted IL-12p40 to a similar extent as DCs previously stimulated with cytokines, while IL-12p70 was not induced in identically treated human DCs. The p38 MAPK was involved in this process, suggesting that the lack of its phosphorylation could be overcome by CD40 ligation. Further analysis revealed that both subsets of DCs also secreted considerable amounts of IL-23, which had only poorly been produced by cytokine-activated DCs after the first activation. Notably, ligation of TLR4 by LPS did not restore the IL-12 production in DCs preactivated for 24 h with GA and d-cAMP. Thus, subsequent T cell stimuli may considerably alter the functions of preactivated DCs. Similarly, Lehner et al. (65) have recently observed an altered cytokine secretion pattern by DCs activated for 48 h with proinflammatory cytokines and restimulated with CD154 and IFN-γ. IL-23 production after stimulation through CD40 has also been shown for mouse DCs (66). Likewise, IL-23 is induced in vivo in T and B cell-deficient mice through CD40 ligation (67). The findings of the present study may have potential implications for the use of cytokine-activated DCs in clinical studies, as the outcome of these studies may also be influenced by cytokines (e.g., IL-23) produced by the cells following restimulation.
Intradermally injected GA/d-cAMP-stimulated monocyte-derived autologous DCs migrated to the T cell areas of draining lymph nodes, where they were located as mature cells in proximity to resident DCs. We have previously observed a similar pattern of migrated rhesus cytokine-activated DCs following s.c. injection (28). While PGE2 is essential for migration of monocyte-derived DCs (68, 69, 70)—and we exposed only cytokine-activated DCs to PGE2—this property of GA/d-cAMP-activated DCs is most likely due to an autocrine effect of PGE2 secretion induced by d-cAMP (71).
We performed a pilot study on the induction of protein-specific immune responses by autologous GA/d-cAMP- or cytokine-activated DCs with the limited number of monkeys available. Notably, all animals responded with similar immune responses to injections with Ag-loaded GA/d-cAMP-activated or cytokine-treated DCs. The induced KLH-specific cytokine secretion patterns in the animals were uniform and characterized by production of IFN-γ, TNF, and IL-17. Secretion of IFN-γ is indicating the induction of Th1 cells. There is evidence for a less essential role of IL-12 for the induction of Th1 immune responses in primates than in rodents. In contrast to IL-12p40 knock-out mice that are extremely susceptible to many infectious agents, there seems to be a redundancy of IL-12 in protection against a large number of microorganisms in humans (72). Thus, similar to the stimulation of Th1 recall responses in vitro, other cytokines as well as chemokines, such as the CCL3 detected in the present study, or cytokines/chemokines produced upon restimulation in vivo (e.g., IL-23) might have contributed to the induction of Th1 immune responses in vivo. Whereas mouse naive T cells cannot be stimulated with IL-23 alone, this cytokine has been shown to activate human CD45RA+ T cells (73). Notably, murine IL-23 can drive the induction of IFN-γ-producing CD4+ T cells in vivo (74). Likewise, IL-12p40 knock-out mice immunized with DNA/IL-23 plasmids, which induced the production of IL-23 but not of IL-12p70 in transfected IL-12p40 knock-out DCs, developed strong Th1 immune responses, rendering the mice partially protected against M. tuberculosis infection (75). Moreover, infection of human DCs with Bordetella pertussis inhibits IL-12p35 production through cAMP induction. Such DCs produce IL-23 but no IL-12p70, but they drive Th1 responses in vitro (76, 77). Thus, the production of IL-23 might also affect the induction of Th cell responses in nonhuman primates and may have influenced, in the context of other yet unidentified chemokines/cytokines and stimulatory membrane-bound molecules, the quality of the immune response induced by the ex vivo-activated DCs.
IL-23 is a key cytokine involved in the development of IL-17-secreting Th cells, and IL-17 was also secreted by the CD4+ T cells obtained from the injected animals. While IL-17 secretion is inhibited by IL-12 in vitro (78), we did not observe differences in IL-17 secretion between the two groups of animals. This may be due to the lack of IL-12p70 production in vivo by either subset of DCs injected. Simultaneous IL-12 and IL-23 secretion, however, leading to the expression of IFN-γ and IL-17 occurs also in experimental autoimmune encephalomyelitis (79). In mice, IL-2 inhibits the generation of Th17 cells (80). In contrast, human Th17 cells can be expanded by IL-2, and neutralizing IL-2 and its receptor inhibits their generation in vitro (81). Thus, it is unlikely that the use of IL-2 to generate CD4+ T cell lines in our study led to an underestimation of IL-17 production by these cell lines. Both Th1 and Th17 cells can produce TNF, the third cytokine produced by the T cells in our experiments. In contrast, IL-4 and IL-10 were not produced. These findings indicate the induction of mixed Th1/Th17 immune responses in vivo irrespective of the DC maturation stimulus used in vitro. Due to the study design we cannot ascertain whether IFN-γ and IL-17 were produced by a single Th or two different Th cell populations.
Both GA/d-cAMP- and cytokine-activated DCs transiently expanded peripheral FOXP3+ regulatory T cells to a similar extent as, or more than, observed with reinjected human cytokine-activated DCs in myeloma patients (82). Whether these cells affected either the magnitude or function of the simultaneously induced KLH-specific T effector cells or both remains to be studied. Notably, the T cells in our in vitro assays did not produce IL-10, which is involved in the regulation of immunity by Ag-induced regulatory T cells (83).
In conclusion, previously activated primate DCs may produce qualitatively different cytokines when these cells additionally receive subsequent stimuli. This may have a substantial impact on strategies of immunotherapy on the basis of reinjected Ag-loaded DCs. The present work emphasizes the need for further studies on DC biology in monkeys to elucidate the underlying mechanisms of the induction of systemic and mucosal cellular immune responses in primates.
We thank Dr. Ralph M. Steinman for helpful discussions and critical reading of the manuscript. We are grateful to Gudrun Groβschupff, Pablo Renner Viveros, Ursula Rüschendorf, Petra Huck, and Maryvonne Paris for excellent technical assistance.
The authors have no financial conflicts of interest.
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 grants from the German Research Foundation (Klinische Forschergruppe 104 to R.I. and T.S., and IG 14/3-1 to R.I.), by the European Union Grant 018685 to K.T.R., C.S.H., P.R., K.Ü., and R.I., and the H.W. & J. Hector Foundation (to R.I.).
Abbreviations used in this paper: DC, dendritic cell; BCG, Mycobacterium bovis bacillus Calmette-Guérin; CMFDA, 5-chloromethylfluorescein diacetate; d-cAMP, dibutyryl-cAMP; GA, glatiramer acetate; KLH, keyhole limpet hemocyanin; PPD, purified protein derivative of Mycobacterium tuberculosis; SEB, staphylococcal enterotoxin B.