SEARCH

SEARCH BY CITATION

Keywords:

  • adenosine;
  • dendritic cells;
  • immunomodulation;
  • T-cell priming

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

Multiple endogenous mechanisms that regulate immune and inflammatory processes contribute to the maintenance of peripheral tolerance and prevent chronic inflammation in mammals. Yet pathogens and tumours are able to exploit these homeostatic pathways to foster immunosuppressive microenvironments and evade immune surveillance. The release of adenosine in the extracellular space contributes to these phenomena by exerting a broad range of immunomodulatory effects. Here we document the influence of adenosine receptor triggering on human dendritic cell differentiation and functions. We show that the expression of several immunomodulatory proteins and myeloid/monocytic lineage markers was affected by adenosine receptors and the cAMP pathway. These changes were reminiscent of the phenotype associated with tolerogenic dendritic cells and, functionally, translated into a defective capacity to prime CD8+ T-cells with a common tumour antigen in vitro. These results establish a novel mechanism by which adenosine hampers CD8+ T-cell immunity via dendritic cells that may contribute to peripheral tolerance as well as to the establishment of immunosuppressive microenvironments relevant to tumour biology.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

Dendritic cells (DCs) are major orchestrators of the immune system that contribute to: (i) maintenance of steady-state tolerance,[1] (ii) initiation of adaptive immune responses, and (iii) development of immune memory.[2-4] In steady-state conditions fully differentiated immature DC (iDC) subsets derive either from blood-borne or tissue-resident precursors and are replenished at different rates.[5] In these conditions monocytes are thought to be a minor source of DCs. Yet in an inflammatory context monocytes differentiate into discrete DC populations that greatly influence ongoing immune responses through the capture and presentation of antigens, the secretion of cytokines and the expression of co-stimulatory molecules.[6, 7] This developmental path can be recapitulated in vitro by driving the differentiation of monocytes into DCs with cytokines, most commonly granulocyte–macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4),[8] resulting in the expression of characteristic DC lineage markers such as CD1a and DC-SIGN and the down-regulation of monocytic markers. The developmental plasticity and functional heterogeneity inherent in DCs is well illustrated by the varied influences that diverse molecules such as metabolites, hormones or cytokines can have on monocyte-derived DC (moDC) differentiation and functions.[9-11]

The nucleoside adenosine is an endogenous molecule that exerts such modulatory activity on moDCs. Extra-cellular adenosine is globally thought of as an endogenous mediator of tissue protection, repair and regeneration that elicits broad modification of cellular metabolism, promotes tissue vascularization and inhibits inflammatory processes.[12, 13] The effects of extracellular adenosine are mediated by the four members of the P1 purinergic G-protein coupled receptor (GPCR) family (A1, A2A, A2B and A3), whose activity results mainly from the modulation of the intra-cellular second messenger cAMP.[14] A2A and A2B are coupled to Gs sub-units and their ligation results in the stimulation of adenylate cyclase and subsequent increases in intra-cellular cAMP, whereas A1 and A3 are coupled to Gi sub-units, which have an opposing effect.[14] The adenosine receptor (AR) 2A appears to be dominant in the immune system[15] and its deletion in mice results in increased systemic concentrations of pro-inflammatory cytokines and tissue damage in concanavalin A-induced and lipopolysaccharide (LPS) -induced models of inflammation.[16] Numerous studies have since reported that adenosine exerts a broad range of effects that combine to down-regulate the immune response and limit the extent of collateral tissue damage occurring during inflammation.[17, 18]

Adenosine is generated in the extra-cellular space through the lysis of ATP, catalysed by ecto-nucleotidases and 5′-nucleotidases expressed on the surface of different cell types, as well as through the inhibition of adenosine phosphorylation[19] and cellular uptake.[20] The principal environmental factor contributing to the increase of extra-cellular adenosine concentration is hypoxia. Local oxygen deprivation occurs during tissue ischaemia, damage resulting from acute inflammation or at the close proximity of rapidly growing solid tumours.[21] The release of adenosine in the tumour environment contributes both to immune evasion and to the direct promotion of tumour growth and vascularization.[21] Moreover, several epithelial tumours can express functional CD39 ecto-nucleotidase and/or CD73 5′-nucleotidase[22] and are therefore able to directly generate extra-cellular adenosine in their vicinity independently of a hypoxic environment. From the standpoint of anti-infectious immunity, genetic deletion of the A2AR resulted in the enhancement of anti-bacterial immunity and in increased survival in a model of polymicrobial sepsis.[23] Importantly, Thammavongsa et al. [24] have shown that preventing the generation of extra-cellular adenosine by Staphylococcus aureus through the genetic deletion of a dedicated microbial nucleotidase increased anti-bacterial immunity and greatly reduced the virulence of this pathogen. The fact that several Gram-negative and Gram-positive bacteria express related enzymes with similar ecto-nucleotidase activity raises the possibility that adenosine generation is a major determinant of bacterial virulence that acts by subverting the immune response. In addition, recent reports showed that adenosine is present at high concentrations (~ 20–100 μm) in the saliva of blood-sucking arthropods and contributes to its strong immunosuppressive and anti-inflammatory properties, notably on DCs.[25, 26] Finally, adenosine is present at high levels in neonatal blood where it has been linked to a decreased production of pro-inflammatory cytokines by monocytes.[27]

Adenosine receptor agonists have a net anti-inflammatory effect on lymphocyte, monocyte, macrophage, DC and neutrophil activities.[17, 18] The relevant mechanism of action is probably best characterized in natural killer cells[28] and T-cells,[29] where it is clearly linked to the A2AR-dependent increase in cytosolic cAMP. Adenosine and its analogues exert anti-inflammatory effects on human moDC[30] and plasmacytoid DC.[31] Functionally AR ligation during moDC maturation alters CD4+ T-cell priming by preventing T helper type 1 (Th1) differentiation in allogeneic mixed lymphocyte reactions.[32] In addition to DC maturation, AR agonists have a disruptive effect on the differentiation of monocytes into DCs. It is unclear, however, how differentiation is affected and what mechanisms account for AR immunomodulatory effects. Here we investigate the influence of the adenosine analogue N-ethylcarboxamidoadenosine (NECA) on moDC differentiation and functions. We report that AR ligation during moDC differentiation results in the acquisition of a phenotype reminiscent of tolerogenic DCs. The secretion of cytokines by these cells upon LPS maturation is also greatly altered and skewed towards an anti-inflammatory profile. Lastly, moDCs differentiated in the presence of NECA are defective at priming CD8+ antigen responses in vitro. All of these features were replicated when using cAMP-elevating agents, suggesting that this signalling pathway is responsible for the effects of adenosine in this context. Altogether our results provide insights into the immunomodulatory effects of adenosine, as well as other cAMP-elevating cues, and point to a novel suppressive mechanism of these pathways on CD8+ T-cell priming by DCs.

Material and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

Monocyte purification and cell culture conditions

Isolation and use of healthy human monocytes was approved by the Cardiff University School of Medicine Research Ethics Committee or the Merck Serono Ethics committee, where applicable. Blood obtained from healthy donors was used to isolate peripheral blood mononuclear cells by gradient centrifugation using Ficoll-plaque plus (GE Healthcare Bio-sciences AB, Uppsala, Sweden). Monocytes were purified by negative magnetic selection with the Monocyte isolation kit II (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). Cells were cultured in RPMI-1640 (Life Technologies, Paisley, UK) containing 100 U/ml penicillin (Life Technologies), 100 mg/ml streptomycin (Life Technologies), 2 mm l-glutamine (Life Technologies), and 10% heat-inactivated fetal calf serum (Life Technologies). Monocytes were induced to differentiate into iDCs for 6–8 days with 100 ng/ml GM-CSF and 50 ng/ml IL-4 (Immunotools, Friesoythe, Germany). For LPS-maturation experiments 50 000 iDCs were washed thoroughly, treated with LPS (100 ng/ml final) in 100 μl of supplemented RPMI-1640 medium and incubated for 24 hr at 37°. The CD8+ T-cell clone Mel5 specific for the HLA A*0201-restricted petide ELAGIGILTV (residues 26–35)[33] derived from Melan-A/Mart1 was used in stimulation assays as described below. Cells were maintained in supplemented RPMI-1640 containing 2·5% Cellkines (Helvetica Healthcare, Geneva, Switzerland), 20 IU/ml IL-2 and 25 ng/ml IL-15 (PeproTech EC, London, UK).

Pharmacological compounds

The non-selective adenosine receptor agonist NECA (Tocris Biosciences, Ellisville, MO), reconstituted in DMSO at 100 mm, was added to a final concentration of 50 μm in the culture medium and replenished every 48 hr. Forskolin (FSK) and 1-methyl-3-isobutylxanthine (IBMX) were purchased from Sigma-Aldrich (Saint Louis, MO) and were stored in DMSO at 50 mm and 400 mm and used at final concentrations of 25 μm and 50 μm, respectively. Experimental conditions referred to in the text as vehicle control were performed with amounts of DMSO corresponding to solvent concentrations determined by the dilution of NECA or FSK and IBMX.

Flow cytometry and antibodies

The moDCs were stained with the viability probe Aqua Dye (Molecular Probes, Paisley, UK) and the following antibodies: anti-CD123-phycoerythrin (PE)-Cy5 (BD Biosciences, Oxford, UK), anti-ILT-3-PE (eBioscience, Hatfield, UK), anti-ILT-4-FITC (eBioscience), anti-CD95-allophycocyanin (APC) (BD Biosciences), anti-CD25-PE (BD Biosciences), anti-CD4-PE-Cy5 (BD-Pharmingen, San Diego, CA), anti-CD325-PE (eBioscience), anti-CD14-FITC, anti-CD1a-AlexaFluor 647 and anti-CD209-FITC (eBioscience). Biotinylated anti-stem cell factor (SCF) antibody was purchased from Biolegend (San Diego, CA). Staining was performed in two steps using either the anti-SCF antibody or an appropriate isotype control first and streptavidin-peridinin chlorophyll protein as a secondary reagent. CD8+ T-cells were stained with HLA A*0201/Melan-A26–35 tetramers (5 μg/ml), anti-CD8-APC and Aqua viability dye. Data acquisition was performed using a FACScalibur (Becton-Dickinson, Oxford, UK) or a FACS Canto II (Becton-Dickinson, UK) and analysis was carried out with FlowJo software (Tree Star Inc., Ashland, OR).

Measurement of cytokine concentration by ELISA, cytokine bead array or Luminex

Interleukin-12 concentration was determined by flow cytometry using the cytometric bead array Human IL-12p70 Flex Set (Becton-Dickinson, San-Diego, CA). Concentrations of IL-23 and IL-27 were determined with the appropriate IL-23 and IL-27p28 Quantikine kits according to the manufacturer's instructions (RnD Systems, Abingdon, UK). The Bio-plex Human Cytokine 27-plex kit (Bio-Rad Laboratories, Hemel Hempstead, UK) was used to measure the concentration of the other reported cytokines with a Luminex 100 instrument (Luminex, Austin, TX).

Autologous CD8+ T-cell clone activation and priming assays

Activation assays were performed in 96-well plates; 15 000 moDCs were matured with 100 ng/ml LPS and pulsed with 10−7 m of agonist peptide for the last 2 hr. Peptide was removed by washing the cells and 200 000 CD8+ T-cell clones specific for the tumour-associated epitope HLA A*0201 Melan-A26–35ELAGIGILTV (EMC Microcollections, Tubingen, Germany), derived from the same healthy donor as the moDCs, were added in presence of 5 μg/ml brefeldin A and anti-CD107a FITC antibody (BD-Pharmingen). Incubation was carried out for 4 hr. Cells were then stained for the surface markers CD8-APC, CD3-PE and the viability probe Aqua Dye (Molecular Probes). Intra-cellular cytokine staining was then performed following fixation and permeabilization of the cells using anti-interferon- γ (IFN-γ) antibodies conjugated to PE-Cy7. For priming assays 200 000 moDCs from HLA A*0201-positive healthy donors were LPS-matured for 48 hr and pulsed with 10−7 m Melan-A26–35 peptide for the last 2 hr. Then, 3 × 106 live autologous total CD8+ T-cells purified by positive selection with magnetic beads (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany) were added per well. Co-cultures were left for 8 days without adding any exogenous factors.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

Changes in the expression pattern of differentiation and lineage markers on moDCs differentiated in presence of AR agonist and cAMP-elevating agents

It was previously reported that treatment with the non-specific synthetic AR agonist NECA during monocyte differentiation into iDCs resulted in high CD14 expression levels and in low CD1a expression levels.[34] We observed similar expression patterns of these molecules in our culture system because moDCs differentiated with vehicle control showed a standard CD14 CD1ahi phenotype (Fig. 1a) whereas in presence of NECA most cells were CD14+ CD1alow (Fig. 1b). In addition, the cAMP-elevating agents FSK and IBMX had a similar effect to NECA on the phenotype of moDCs (Fig. 1c). We investigated whether the expression of other differentiation and lineage markers was affected by treatment with NECA or FSK and IBMX and found that the lymphocytic/monocytic marker CD4, as well as the early haematopoiesis and myelopoiesis markers SCF (KIT ligand) and CD123 (the α subunit of the IL-3 receptor), were expressed on NECA- or FSK/IBMX-treated cells but absent in bona fide moDCs (Fig. 1d–f). In addition, we also observed that the tumour necrosis factor receptor (TNFR) super-family member Fas (CD95) was down-regulated on the cell surface when NECA or FSK/IBMX were present in the differentiation medium (Fig. 1g).

image

Figure 1. Differential expression of lineage and differentiation markers on monocyte-derived dendritic cells (moDCs) differentiated in the presence of vehicle, N-ethylcarboxamidoadenosine (NECA) or forskolin/1-methyl-3-isobutylxanthine (FSK/IBMX). Typical differentiation status determined by monitoring CD14 and CD1a expression on moDCs generated with granulocyte–macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) in the presence of vehicle control (a), NECA (b) or FSK and IBMX (c). Cell numbers in each quadrant are shown as insets in each dot plot. The total numbers of acquired events were 10 080 (a), 12 090 (b) and 13 425 (c). The expression levels of three different myeloid and monocytic markers (d–f), as well as CD95/Fas (g), were differentially regulated by differentiation in presence of NECA or FSK and IBMX. The moDCs were stained with the indicated antibodies after 7 days of differentiation. Data are representative of at least five experiments for each staining.

Download figure to PowerPoint

Several immunomodulatory molecules are up-regulated on the surface of moDCs differentiated with AR agonist or cAMP-elevating agents

Adenosine is believed to exert broad immune modulatory functions in the steady-state, during Th2-type inflammation and in the tumour microenvironment. We therefore sought to determine whether moDC differentiation in the presence of AR agonist would result in the expression of proteins with known regulatory functions. Flow cytometry analysis showed an increase in CD25 (the protein product of IL2RA) expression on the cell surface (Fig. 2a) but not of the secreted soluble form (data not shown). Moreover, the inhibitory receptors ILT-3 and ILT-4 were up-regulated on the surface of moDC NECA and moDC FSK/IBMX (Fig. 2b,c). These data show that non-selective stimulation of ARs by NECA results in the expression of immunosuppressive molecules on the surface of differentiating moDCs, most likely in a cAMP-dependent manner.

image

Figure 2. Up-regulation of proteins involved in immune regulatory mechanisms on the surface of monocyte-derived dendritic cells (moDCs) differentiated in the presence of N-ethylcarboxamidoadenosine (NECA) or forskolin/1-methyl-3-isobutylxanthine (FSK/IBMX). The moDCs were stained with the indicated antibodies after 7 days of differentiation in the presence of granulocyte–macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) with vehicle control, NECA or FSK/IBMX, as indicated in the key (a–c). Data are representative of at least six experiments for each staining.

Download figure to PowerPoint

The alternate differentiation of moDCs impacts their cytokine profile upon LPS maturation

It was previously reported that the presence of AR agonists during DC maturation inhibited the secretion of IL-12.[30, 32] In addition, we found that the secretion of the IL-12 family cytokines IL-23 and IL-27 was also greatly reduced following LPS treatment of moDCs differentiated with NECA or FSK/IBMX (Fig. 3a), even though the compounds were absent during maturation with LPS. We also found that moDCs differentiated with FSK/IBMX released significantly higher amounts of the cytokine granulocyte colony-stimulating factor (G-CSF) and of the decoy receptor IL-1Ra (Fig. 3b). A similar trend was observed for moDCs differentiated with NECA, yet comparison of sample groups was not statistically significant (= 0·0782 for IL1Ra and = 0·0706 for G-CSF). The release of TNF-α and IL-1β was down-regulated whereas the concentrations of IL-10 and vascular endothelial growth factor were substantially increased upon LPS stimulation (Fig. 3b), as previously reported.[32, 34] No other difference was observed among other factors secreted by moDCs (Fig. 3c). Overall, this shift in cytokine secretion profile shows a skewing towards an anti-inflammatory activity and suggests alterations of the T-cell stimulatory properties in moDCs differentiated with NECA and FSK/IBMX.

image

Figure 3. Comparison of the cytokine profiles of monocyte-derived dendritic cells (mo-DCs) differentiated in the presence of N-ethylcarboxamidoadenosine (NECA) or vehicle control upon lipopolysaccharide (LPS) maturation. The moDCs were differentiated with granulocyte–macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) for 6–8 days with vehicle control, in the presence of NECA or forskolin/1-methyl-3-isobutylxanthine (FSK/IBMX). Cells were then treated with LPS for 24 hr in the absence of any other compound. (a) Cytokine concentrations in supernatants were determined by cytometric bead arrays and ELISA for the IL-12 cytokine family or (b) and (c) by using the 27-plex Luminex bead combination. Assays were performed in triplicate for six donors in the case of IL-12, IL-23 and IL-27 ELISAs and for three donors in the case of the 27-plex Luminex bead arrays. Representative data are shown. P values were calculated using the two-tailed paired student t test (*< 0·05; **< 0·01; ***< 0·001; ns: non-significant).

Download figure to PowerPoint

Deficient in vitro priming of antigen-specific CD8+ T-cell responses by mature DCs differentiated with NECA or FSK and IBMX

We next assessed CD8+ T-cell activation by moDCs differentiated with AR agonists or cAMP-elevating agents, first against long-term CD8+ T-cell clones and, second, in priming experiments using autologous peripheral blood CD8+ T-cells. In short intra-cellular cytokine staining assays moDC NECA and moDC FSK/IBMX elicited similar antigen-specific cytotoxic granule exocytosis and IFN-γ synthesis levels as moDCs (Fig. 4a), indicating that these cells exerted no suppressive activity on highly differentiated effector CD8+ T-cells maintained with homeostatic and differentiation cytokines. This result also suggested that all three moDC types express similar levels of agonist peptide major histocompatibility class I (pMHCI) molecules when pulsed with identical concentrations of exogenous peptide. Next we compared the capacity of each type of moDCs to prime a response against the HLA A*0201 restricted Melan-A26–35 peptide following LPS maturation. pMHCI recognition dominated by TCR germline amino acid residues[33] ensures high frequencies (≥ 0·05%) of CD8+ T-cells specific for this epitope in healthy naive HLA A*0201-positive individuals,[35] making it an ideal system to assess human antigen-specific responses in vitro. We observed high frequencies of antigen-specific CD8+ T-cells, as measured by tetramer staining, with bona fide mature moDCs (Fig. 4b,c). Strikingly, priming experiments performed with moDCs generated in the presence of NECA or FSK/IBMX significantly induced much lower expansions of antigen-specific CD8+ T-cells (Fig. 4b,c), indicating that these cells are impaired in their ability to initiate CD8+ T-cell responses.

image

Figure 4. Assessment of the stimulatory and CD8+ T-cell priming capacities of mature monocyte-derived dendritic cells (moDCs) differentiated in the presence of N-ethylcarboxamidoadenosine (NECA) or forskolin/1-methyl-3-isobutylxanthine (FSK/IBMX) compared with bona fide moDCs. (a) The CD8+ T-cell clone Mel5 was stimulated in intra-cellular cytokine staining assays, as described in the Materials and methods section, using mature moDCs without peptide or moDCs, moDC NECA and moDC FSK/IBMX pulsed with agonist peptide as indicated. Staining is representative of seven different experiments. (b) Proportions of antigen-specific CD8+ T-cells following an 8-day priming experiment using peptide-pulsed and lipopolysaccharide (LPS) -matured moDCs, moDC NECA or moDC FSK/IBMX as antigen-presenting cells. The percentage of tetramer-positive cells is indicated in each plot for cells gated as follows: Forward-scatter (FSC-H) and side-scatter (SSC-H) lymphocyte gate/Singlet determined by FCS-A/FSC-H analysis/Live CD8+ cells negative for the Aqua viability dye. The baseline frequency of tetramer-positive CD8+ cells, assessed in a staining performed before priming, was below 0·1% (data not shown). Absolute cell counts were as follows: DCs 0 (85 726 gated events and 8144 tetramer-positive cells); NECA DCs (81 731 gated events and 1365 tetramer-positive cells); and F/I DCs (71 826 acquired events and 232 tetramer-positive cells). (c) Comparison of the priming capacities of mature, peptide-pulsed moDCs, moDC NECA or moDC FSK/IBMX (as indicated) for seven different donors. Bars represent mean values (11·23% for DCs 0, 2·21% for NECA DCs and 0·75% for F/I DCs). The ranges of tetramer-positive frequency values were: 3·3–29·6% for moDCs, 0·7–4·35% for moDC NECA and 0·32–1·1% for moDC FSK/IBMX. P-values were calculated using the two-tailed unpaired Student's t-test.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

In this study we report the effects of an AR agonist on the phenotype and functions of DCs differentiated from monocyte precursors. We correlated changes in the expression of several molecules involved in immune, apoptotic and developmental processes with the presence of the synthetic adenosine analogue NECA or of the cAMP-elevating agents FSK and IBMX during differentiation. This altered differentiation state also had a dramatic impact on the cytokine secretion pattern and the immunogenicity of LPS-matured DCs. Cells differentiated in the presence of NECA or FSK/IBMX showed an anti-inflammatory cytokine profile and were defective with respect to the priming of an antigen-specific CD8+ T-cell response against the HLA A*0201 restricted tumour-associated antigen Melan-A26–35 in vitro.

The release of extra-cellular adenosine is widely regarded as an endogenous mechanism of tissue repair and regeneration that mediates protective effects on the organism at multiple levels, including extensive regulation of the immune system. We identified regulatory pathways in iDCs exposed to NECA that fit this concept. Notably, we showed an up-regulation of the high-affinity IL-2R sub-unit CD25 (Fig. 2a). Human DCs do not express a functional IL-2R because of lack of expression of the β sub-unit and CD25 is therefore believed to act as a decoy receptor, decreasing the availability of IL-2 for activated T-cells.[36] So far, CD25 expression has been linked to DC maturation cues, e.g. TLR agonists or pro-inflammatory cytokines.[36] Our results show that CD25 expression can also be induced in iDCs by an adenosine analogue and, more broadly, by signals that trigger intra-cellular cAMP up-regulation. In addition, the inhibitory receptors ILT-3 and ILT-4, known to exert immunoregulatory functions on T-cell stimulation by DCs,[37] were also expressed on the cell surface as a consequence of AR ligation by NECA or in the presence of FSK/IBMX (Fig. 2b,c). Interestingly, the expression of membrane-bound SCF was also affected (Fig. 1e). In addition to its crucial role in haematopoiesis the expression of SCF and of its receptor c-KIT (CD117) can be induced by cAMP-inducing signals in murine DCs.[38] In that study the autocrine activity of membrane and soluble SCF resulted in skewing the polarization of CD4 T-cells towards Th2 and Th17 differentiation. Another differentially regulated gene product relevant to DC biology is the TNFR super-family member Fas/CD95. Fas-mediated apoptosis is believed to be the main mechanism of DC death both in the steady-state and during the course of immune responses.[39, 40] Differentiation in the presence of NECA and, more strikingly, in the presence of FSK and IBMX decreased the expression of Fas, implying that DCs exposed to cAMP-elevating agents in general and to AR agonists in particular may be less prone to Fas-induced apoptosis. In addition to the regulation of cell-surface proteins, moDCs differentiated with NECA or FSK/IBMX exhibited an altered cytokine profile upon LPS maturation, even though the compounds were absent during LPS treatment. The most striking feature was the almost complete inhibition of secretion of the IL-12 family cytokines. Interference with IL-12p70 secretion by AR agonists had already been described but inhibition of both IL-23, important for the differentiation and/or expansion of human Th17 cells,[41] and IL-27 are novel features. These results suggest additional mechanisms through which DCs differentiated with AR agonists might regulate the T-cell response. The increased secretion of G-CSF and IL-1Ra are also new features revealed by our study.

The presence of molecules with immunoregulatory functions on moDC NECA, as well as their cytokine profile upon maturation, are reminiscent of the descriptions made for subsets of DCs referred to in the literature as regulatory DCs (DCreg) or tolerogenic DCs (tolDCs). These tolDCs are broadly defined as DCs that exert regulatory functions on the immune system and they are believed to make an important contribution to the maintenance of immune tolerance in the steady-state.[42, 43] Hallmark phenotypic markers of this DC subset are the enzyme indoleamine 2,3-dioxygenase (IDO)[44] and the CD25 molecule,[45] both up-regulated in moDC NECA as shown by Novitskiy et al.[34] and us (Fig. 2a), respectively. Other features we identified in moDC NECA; such as high-level expression of CD123, IL-10 secretion or IL-12 inhibition, match the phenotypes commonly assigned to subsets of tolDCs.[44, 46] At present it is unclear what mechanisms account for the suppressive activity of tolDCs. Our study reveals the coordinated expression of numerous molecules directly relevant to immunosuppressive activities in DCs that may take part in the tolerogenic properties of moDC NECA and of naturally occurring or induced tolDC populations. Functionally tolDCs exert a suppressive activity on T-cell activation both in vitro and in vivo.[47] MoDC NECA skew CD4 T-cell responses towards a Th2 profile in in vitro allogeneic mixed lymphocyte reactions,[32, 34] a hallmark functional feature of tolDCs. We show here that LPS-matured DCs generated in the presence of the non-specific adenosine analogue NECA or of cAMP-elevating agents fail to prime CD8+ T-cell responses in vitro. This result is particularly relevant because, in addition to its protective effects on tissue homeostasis, recent reports point to the involvement of adenosine release as a strategy evolved by both solid tumours and pathogens to subvert the immune response.[24, 48] CD8+ T-cell immunity confers efficient protection against intra-cellular pathogens and tumours; the failure of its initiation by means of subversion of innate immune mechanisms would represent an efficient escape strategy. It is noteworthy that all the phenotypic features we uncovered in moDC NECA could be replicated by increasing intra-cellular cAMP with FSK and IBMX. Many signals up-regulating the synthesis of cAMP, such as those provided by ligands of Gs-coupled GPCRs, are therefore likely to induce moDCs with similar properties. For instance the addition of prostaglandin E2 or vasointestinal peptides to the culture media of differentiating moDCs resulted in the acquisition of a phenotype similar to that of moDC NECAs.[9, 45] Of note Kleijwegt et al.[49] have shown in a recent report that tolDCs generated from human monocytes with a combination of dexamethasone and vitamin D3 completely failed to activate naive CD8+ T-cells and elicited sub-optimal memory responses. Our experimental system does not allow distinction to be made between these two subsets because we used total blood CD8+ T-cells. However, a similar scenario appears plausible for moDC NECA because the proliferation of antigen-specific CD8+ T-cells is highly decreased but not completely impaired.

In summary, we have shown that ARs and the cAMP second messenger pathway differentially regulate the expression of several key immune molecules in DCs, resulting in a phenotype compatible with the regulation of inflammation and immunity. These findings extend the range of putative immunosuppressive mechanisms used by DCs differentiated in the presence of adenosine, of cAMP-elevating agents and possibly by several subsets of so-called tolDCs. In addition to immunosuppressive proteins we found that several developmental molecules and lineage markers were regulated by NECA and by FSK/IBMX. In light of our results it is conceivable that the phenotype of moDC NECA and of tolerogenic moDCs in general reflect the fundamental developmental plasticity of DCs, illustrating the influence that extrinsic factors from the extra-cellular milieu can exert on the phenotype and functions of DCs, and pointing to immune evasion strategies relying on the induction of tolDCs with respect to the initiation of CD8+ T-cell immunity.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

BL and AKS are funded by the Biotechnological and Biological Sciences Research Council (UK) grant BB/H001085.

The authors wish to thank Dr Fabien Blanchet for helpful comments and critical reviewing of the manuscript.

Disclosures

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

DB and JC are employees of Merck Serono S.A.-Geneva, an affiliation of Merck KGaA, Darmstadt, Germany. BL is a former employee of Merck Serono S.A.-Geneva.

Authorship declaration

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References

BL designed the study and wrote the article. AKS contributed to the preparation of the manuscript. BL and JC performed experiments. BL, DB and AKS analysed the data.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. Authorship declaration
  10. References