Recently it has been shown that dendritic cells (DC) express both Notch and Notch ligands, allowing for the possibility that Notch signaling may influence their maturation. We show that although both Jagged (Jgd) and Delta-like (DlL) ligands were able to activate the canonical Notch pathway in mouse DC, only Jgd1 could induce the production of certain cytokines. Maturation of DC via Jgd1 resulted in an entirely different maturation program from that induced through TLR (via LPS) signaling, promoting the production of high levels of IL-2 and IL-10. DC matured by Jgd1 (Jgd1-conditioned DC) promoted the survival and proliferation of CD4+CD25+ regulatory T cells that were able to suppress efficiently the proliferation of CD25− cells. Further, CD25+ cells cultured with Jgd1-conditioned DC produced very high levels of IL-17 in an IL-2-dependent fashion. Our data suggest a new and important role for the Notch pathway in the regulation of the DC phenotype.
Dendritic cells (DC)3 play a central role in the innate and adaptive immune system. Scattered throughout the body, they act as sentinels for the detection of pathogens. Depending on the environment, the type of signal perceived and their maturation state, they can produce inflammatory (IL-12, IL-6, IL-15, IL-18, TNF) or regulatory (IL-10) cytokines.
It is now well established that activation of the Notch pathway influences immune cell fate. Notch plays a role in development of αβ and CD4/CD8 T cells in the thymus (1, 2) and in T cell differentiation in the periphery (3). In contrast, much less is known about the effects of Notch on the maturation and/or activation of APC, including DC. Both Notch receptors and their ligands are expressed on DC and it has been shown that bacterial products not only up-regulate the expression of Notch ligands in DC (4) but also can activate Notch signaling in macrophages (5). There is little information about the effects of Notch signaling on DC phenotype or function, although one study suggested that Notch stimulation with Jagged (Jgd) 1 could promote monocyte-derived DC to mature in a manner similar to that induced by LPS (6).
In this study, we characterize the influence of Notch signaling, delivered by either Jgd or Delta-like (DlL) ligands, on the phenotype and function of DC.
This study provides new insight into how Notch signaling via Jgd1 can alter DC maturation that in turn can promote the growth of T regulatory cells (Tregs) and the production of IL-17.
Materials and Methods
Male C57BL/6 mice, (8 to 12 wks) (Harlan) were maintained in accordance with U.K. Home Office guidelines (Animals (Scientific Procedures) Act 1986).
Bone marrow-derived DC cultures
Bone marrow-derived DC cultures were prepared from wild-type or IL-2 knockout (KO) (gift from F. Powrie, University of Oxford, Oxford, U.K.) mice as described (7) with some modifications. At days 3 and 6, fresh medium containing 20 ng/ml GM-CSF (B hybridoma (X63) supernatant) was added. DC were used after 8 days.
Immobilized rat Jgd1/humanFc fusion protein, human rDlL4 (R&D Systems), and DlL1/human Fc (Lorantis) were used as Notch receptor ligands. Human IgG1 (Sigma-Aldrich) (control DC) and PBS were used as control. Initial experiments showed that the following concentrations were optimum for the effects observed: Jgd1 (10 μg/ml), DlL1 (50 μg/ml), DlL4 (10 μg/ml). Plates were coated with ligands or IgG1 in PBS overnight at 4°C. LPS was from Escherichia coli 026:B6 (Sigma-Aldrich).
Quantitive RT-PCR (qRT-PCR)
cDNA was generated from 125 ng of total RNA using an archive kit (Applied Biosystems). One μl of cDNA was used with PCR master mix and TaqMan assays (Applied Biosystems). All reactions were conducted in triplicate using an Applied Biosystems 7500 PCR system. Cycle thresholds obtained were normalized to 18S and calibrated to a PBS-treated sample for relative quantification. Arbitrary units relative to detection limit levels were used for IL-12p40 and IL-23p19 mRNA because they were undetectable in PBS controls.
After 24 h of treatment, DC were labeled with allophycocyanin-labeled CD11c, PE-labeled MHCII, CD80, CD86, and CD40 Abs (all from BD Pharmingen). For Foxp3 (eBioscience) staining, cells were first fixed and permeabilized with Perm/Wash buffer (BD Biosciences). For intracellular IL-17 staining, cells were stimulated with PMA/ionomycin and brefeldin A for the last 4 h of coculture. Cells were stained using allophycocyanin-labeled CD11c, Cy5-Alexa Fluor 700-labeled CD4, PE-labeled CD25, FITC-labeled γδ, NK1.1, CD8, or CD3 Abs. Analysis was performed on a four-color FACSCalibur flow cytometer (BD Biosciences).
CD4+CD25+ and CD4+CD25− cell purification, proliferation, and suppression assays
Lymph node CD4+CD25+ and CD4+CD25− T cells were positively and negatively purified, respectively, whereas CD25+ cells were only positively purified, all using a MACS kit (Miltenyi Biotec). Proliferation or cytokine production was assessed by culturing CD4+CD25+ cells with Jgd1-conditioned or control DC for 72 h or 7 days. For IL-2R blocking, p55 Ab (PC61.5, Insight Biotechnology) was used. TGF-β1 Ab was used to neutralize bioactivity (R&D Systems). Suppression assays were set up with purified CD4+CD25− responder T cells in round-bottom wells with 0.2 μg/ml CD3 Ab (clone 145-2C11; BD Pharmingen), irradiated (3,000 rad) splenocytes, and purified CD4+CD25+ fresh or following culture with Jgd1-conditioned or control DC at a 2:1 ratio. Proliferation was assessed by [3H]thymidine incorporation.
Mouse IL-17 (R&D Systems) IL-2, IL-6, and IFN-γ (BD Biosciences) kits were used.
Results and Discussion
Notch signaling initiated by Jgd1, but not DlL, induces a distinctive maturation profile in DC
Stimulation of DC with the Notch ligands Jgd1, DlL1, or DlL4 induced similar levels of Notch signaling as assessed by increased transcripts levels of the Notch target gene, Hes1 (Fig. 1,A). In agreement with others (8), we found that in DC IL-2 transcripts were increased by ∼4-fold in the presence of LPS. When DC were activated by Jgd1 we observed a 50-fold increase in IL-2 transcripts at 4 h (Fig. 1,B), but when cells were stimulated with DlL1 or DlL4, no increase in IL-2 transcripts was detected. IL-2 transcript levels following Jgd1 ligation peaked with 275-fold induction at ∼2.5 h (Fig. 1,C). By contrast, LPS induced 5-fold fewer IL-2 transcripts with a later peak of expression at ∼3 h. IL-2 protein levels (24 h) were also shown to be greater following stimulation of DC with Jgd1 than with LPS (Fig. 1 D). These results suggest that all three ligands were able to stimulate the canonical pathway of Notch signaling but that the increase in IL-2 expression induced by Jgd1 might involve additional components of Notch signaling. A differential effect of the various Notch ligands has been previously observed in T cells where DlL1, DlL4, and Jgd1 differentially regulated activation of peripheral Th cells (9) or precursor T cells (10).
Although both Jgd1 and LPS induced an increase in IL-10 and TNF-α transcripts (Fig. 1, E and F), we found that IL-6, IL-12p40, IL-15, and IL-23p19 were only induced by LPS (Fig. 1, G–J), indicating that LPS and Jgd1 elicited different maturation responses in DC. Neither DlL1 nor DlL4 were able to stimulate expression of any of these cytokines. LPS-induced IL-2 production in DC has been shown to be IL-15 dependent (11). In our experiments, although IL-15 was induced by LPS it was not induced by Jgd1. This, combined with the delayed kinetics of LPS-induced IL-2 expression, suggests that the regulation of IL-2 by LPS and Jgd1 may involve different mechanisms.
Classical cell surface markers of DC maturation were also analyzed. Expression of MHC class II, CD80, and CD86 were increased to the same extent by both LPS and Jgd1 (data not shown), and this is consistent with data previously reported by Weijzen et al. (6). In contrast, CD40 was only significantly up-regulated by LPS. CD11c was slightly increased by Jgd1 ligation but not by LPS stimulation (mean fluorescence intensity of 84 with Jgd1 compared with 41 and 43 in untreated and LPS-treated DC, respectively).
Taken together, our results show that signaling via the endogenous ligand Jgd1 can induce a previously unreported maturation profile in DC. Although Weijzen et al. (6) suggested previously that LPS and Jgd1 could induce similar maturation profiles in DC, they did not analyze the range of cytokines shown here. Only IL-12 was measured and was induced to much lower levels by Jgd1 than by LPS.
Jgd1-conditioned DC support the proliferation and suppressive capacity of CD4+CD25+ Treg
IL-2 is critical to Treg survival and proliferation (12, 13). Indeed, growth of Treg in culture invariably requires the addition of IL-2. To assess whether Jgd1-conditioned DC could affect the survival and/or regulatory capacity of CD4+CD25+ T cells, cocultures of these cells were performed.
DC were activated with Jgd1 for 4 h before the addition of purified CD4+CD25+ T cells. In this syngeneic system with no TCR stimulus being provided, Jgd1-conditioned DC promoted CD4+CD25+ cell survival (Fig. 2,A). After 7 days of culture about three times more T cells were consistently recovered from Jgd1-conditioned DC than from control DC cocultures. This effect was IL-2 dependent, because Ab-mediated blocking of the IL-2R abrogated the increased proliferation of CD4+CD25+ T cells in the Jgd1-conditioned DC cocultures but had no effect on that of CD4+CD25+ T cells cocultured with control-DC (Fig. 2,B). DC from IL-2 KO mice were unable to sustain an increased proliferation of CD4+CD25+ cells (data not shown) and recovery of CD4+CD25+ T cells was similar from cocultures with Jgd1 DC and control DC (Fig. 2 C), showing that it was IL-2 derived from DC that was important for the growth advantage conferred upon the CD4+CD25+ cells in cocultures.
After 7 days of coculture with irradiated, Jgd1-conditioned, or control DC, CD4+CD25+ cells were collected and assessed for their ability to inhibit the proliferation of freshly purified CD4+CD25− T cell responders. CD4+CD25+ T cells cultured with Jgd1-conditioned DC were able to suppress CD4+CD25− T cells as least as efficiently as freshly isolated CD4+CD25+ T cells or Treg cultured on control DC (Fig. 2 D). Statistics could not be performed for assays where CD4+CD25+ cells had been conditioned by control DC, as few cells were recovered and only duplicates were performed. Suppression was confirmed using CFSE-labeled responder CD4+CD25− T cells (data not shown).
These results show that Jgd1-conditioned DC are able to stimulate CD4+CD25+ T cells to proliferate in the absence of added IL-2 or a TCR signal and that these stimulated CD4+CD25+Foxp3+ T cells remain highly suppressive.
Jgd1-conditioned DC stimulate CD25+ lymphocytes to produce high levels of IL-17 in an IL-2-dependent fashion
Because IL-2 is produced by Jgd1-conditioned DC and promotes CD4+CD25+ T cell proliferation, IL-2R α-chain-bearing lymphocytes were assessed for their cytokine production when cocultured with Jgd1-conditioned DC. High levels of IL-17 were found in supernatants of CD25+ lymphocytes cocultured with Jgd1-conditioned DC, whereas no IL-17 was detected in coculture with control DC. Jgd1 had no direct effect on the production of IL-17 by CD25+ cells, and DC did not produce IL-17 themselves (Fig. 3,A). When CD25− cells were used in this coculture system no IL-17 was detected, indicating that the presence of the IL-2R α-chain on the surface of responder cells was essential. Analysis by flow cytometry showed that ∼9% of cells present in coculture were responsible for the high levels of IL-17 production observed (Fig. 3 B).
IL-17 was produced by several immune cell types including γδ T cells, NK cells, and CD8+ and CD4+ T cells, as well as a proportion of Foxp3+ cells (Fig. 3,C). Recently, human CD25highFoxp3+ T cells have also been shown to be able to produce IL-17 when stimulated by allogeneic monocytes and recombinant human (rh)IL-2/rhIL-15, highlighting the great plasticity and diversity of Foxp3+ cells (14). It has been previously shown that IL-17, which regulates granulopoiesis through G-CSF, is made by γδ T cells and unconventional αβ T cells (15). Further, in a model of Mycobacterium tuberculosis infection IL-17 was released predominantly by γδ T and CD4−CD8− cells rather than by CD4+ T cells (16). In those studies, IL-23 produced by DC was shown to be responsible for the production of IL-17, whereas in our system IL-2 is implicated. Indeed, in our system Ab-mediated blocking of the IL2-R completely inhibited IL-17 production (Fig. 3,D), suggesting that IL-17 produced by CD25+ cells cocultured with Jgd1-conditioned DC was entirely IL-2 dependent. IL-17 production was dependent on DC-derived IL-2, because only background levels of IL-17 were detected in cocultures with Jgd1-conditioned DC from IL-2 KO mice (Fig. 3,E). Addition of IL-2 to CD4+ T cells under Th17-polarizing conditions (IL-6 and TGF-β1 in the mouse) has been shown to inhibit differentiation of Th17 cells (17). In contrast, our results show that IL-2 is necessary for the production of IL-17 by CD25+ cells, indicating that IL-2 is likely to play a role in sustaining IL-17 production by T cells. Although IL-6 and TGF-β1 play an important role in Th17 differentiation, they do not appear to be required in the present system because we only detected background levels of IL-6 at the RNA (Fig. 1 G) and protein levels (15 and 50pg/ml for control DC and Jgd1-conditioned DC, respectively, compared with 20,000 pg/ml for LPS-treated DC). When neutralizing TGF-β1 Abs were added to the coculture there was no decrease of IL-17 production induced by Jgd1-conditioned DC on CD25+ cells (data not shown).
Low levels of IFN-γ (up to 2,000 pg/ml) were detected in cocultures of either CD25+ or CD25− cells with either Jgd1-conditioned or control DC. No significant amounts of IL-2 or IL-10 were detected in these cultures above the amounts produced by the DC themselves. These data suggest that the distinct feature of Jgd1 conditioning on DC was restricted to their ability to induce IL-17 production in cocultures with CD25+ cells.
These data show, to our knowledge for the first time, that Notch signaling initiated by Jgd1, but not DlL1 or DlL4, induces a maturation program in DC that is different to that induced by TLR ligands. Although surface expression of MHC, CD80, and CD86 were increased in a similar fashion following both stimuli, cytokines showed a distinct pattern of expression. It has recently been shown that under steady-state conditions DC maturation and migration occurs in the absence of pathogens in either germfree or TRIF/Myd88-deficient mice (18), suggesting that endogenous mediators must provide the maturation signals that play a key role in DC homeostasis. Because Jgd1 is normally expressed on a variety of cells such as keratinocytes, lymphoid tissue itself, and bone marrow, there is a possibility that Notch signaling contributes to this process in the absence of inflammatory signals.
Our data further show that the Notch ligand Jgd1 induces IL-2 production in DC that, in turn, can promote the survival, proliferation, and suppressive abilities of CD4+CD25+ Treg. IL-2 signaling has been shown by others to play an essential role in Treg maintenance in the periphery (12, 13). Our results suggest that via Notch signaling, DC could be a source of IL-2 and promote Treg fitness in tissues such as the intestine. Endogenous signals mediated via Jgd1 could contribute to the homeostasis of DC and consequently to the maintenance and/or expansion of Treg in the absence of TLR or TCR signaling. Conversely, Jgd1 expression associated with tumors, as described for prostate cancer metastases and breast cancer (19, 20), could play a detrimental role by promoting undesirable regulation in the immune system.
CD25+ cells responded to IL-2 signaling by producing IL-17. IL-17 production was mediated by a variety of CD25+ cell types including γδ T cells, NK cells, CD8+, and CD4+Foxp3+ T cells. Our results indicate that Treg cells grown on Jgd1-conditioned DC can still suppress the response of CD4+CD25− responder cells. This was despite the high levels of IL-17 production in cultures, indicating that production of this cytokine alone is not inherently related to an inability to regulate.
Our findings could have important implications in diseases such as multiple sclerosis, where Jgd1 but not DlL1 is specifically re-expressed by hypertrophic astrocytes impeding remyelination (21) and where IL-17 is believed to play a detrimental role. Administration of a Jgd1 fusion protein (22) or peptide (23), which was found to have a beneficial effect overall, could have a number of roles in experimental autoimmune encephalomyelitis, first on remyelination, second on T cells, and finally, as implicated by the results of our study, on DC themselves. Further research will be crucial in determining the exact role of each component of the Notch pathway in this complex system to define its potential as a therapeutic target.
We thank Lorantis for providing us with DlL1-Fc protein and Fiona Powrie for providing the IL-2 KO mouse samples. We thank Dan Davis and Carol Pridgeon for critical reading of the manuscript.
The authors have no financial conflict 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 funded by the Biotechnology and Biological Sciences Research Council and the Wellcome Trust.
Abbreviations used in this paper: DC, dendritic cell; DlL, Delta-like ligand; Jgd, Jagged; KO, knock-out; Treg, T regulatory cell; qRT-PCR, quantitative RT-PCR.