Tolerogenic dendritic cells (DCs) play an important role in maintaining peripheral tolerance through the induction/activation of regulatory T cells (Treg). Endogenous factors contribute to the functional development of tolerogenic DCs. In this report, we present evidence that two known immunosuppressive neuropeptides, the vasoactive intestinal peptide (VIP) and the pituitary adenylate cyclase-activating polypeptide (PACAP), contribute to the development of bone marrow-derived tolerogenic DCs in vitro and in vivo. The VIP/PACAP-generated DCs are CD11clowCD45RBhigh, do not up-regulate CD80, CD86, and CD40 following LPS stimulation, and secrete high amounts of IL-10. The induction of tolerogenic DCs is mediated through the VPAC1 receptor and protein kinase A, and correlates with the inhibition of IκB phosphorylation and of NF-κBp65 nuclear translocation. The VIP/PACAP-generated DCs induce functional Treg in vitro and in vivo. The VIP/DC-induced Treg resemble the previously described Tr1 in terms of phenotype and cytokine profile, suppress primarily Th1 responses including delayed-type hypersensitivity, and transfer suppression to naive hosts. The effect of VIP/PACAP on the DC-Treg axis represents an additional mechanism for their general anti-inflammatory role, particularly in anatomical sites which exhibit immune deviation or privilege.
Dendritic cells (DC)3 are a heterogenous population of APCs that contribute to innate immunity and initiate adaptive immune responses to Ags associated with infection and inflammation (1). The successful initiation of the adaptive immune response requires DC maturation, following signaling through the toll-like receptors and CD40. However, in addition to their proinflammatory role, DC also play an important role in immune homeostasis, by inducing and maintaining tolerance (2).
In addition to the elimination of self-reactive T cells in the thymus, tolerance is maintained in the periphery through clonal deletion, induction of anergy, and differentiation of regulatory T cells (Treg). Several populations of CD4 Treg have been described and characterized, including the naturally occurring CD4+CD25+ Treg (nTreg) and the induced CD4+CD25+ peripheral Treg (iTreg), consisting of IL-10-producing Tr1 and TGF-β-secreting Th3/Tr2 (3, 4, 5, 6).
In contrast to the nTreg generated in the thymus, iTreg are generated in the periphery, and DCs appear to play an essential role in their differentiation. Several recent reports indicate that immature or semimature DCs contribute to the induction of iTreg. Repeated exposure to immature monocyte-derived human DCs secreting IL-10 has been shown to induce IL-10 and TGF-β-secreting CD4 Tr1 (7). Tolerogenic DCs have been generated in the presence of 1,25(OH)2D3, the active form of vitamin D3, and of vitamin D3 analogues from both human blood monocytes and murine bone marrow and shown to induce Treg (8, 9, 10). In addition, DCs with a semimature status were shown to be tolerogenic and induce Tr1 differentiation (11). Although most reports focused on the maturation stage as directly connected to the tolerogenic capacity, a recent study by Wakkach et al. (12) suggests the existence of a distinct subset of tolerogenic DCs characterized by the stable phenotype CD11clowCD45Rbhigh. These DCs can be derived from bone marrow cells in the presence of GM-CSF, TNF, and IL-10, secrete high levels of IL-10 following activation, and induce Tr1 cells in vivo and in vitro. CD11clowCD45RBhigh tolerogenic DCs are present in the spleen and lymph nodes of normal mice, and splenic stromal cells were shown to direct their differentiation (12, 13).
The neuropeptides vasoactive intestinal peptide (VIP) and pituitary adenylate cyclase-activating polypeptide (PACAP) are potent immunosuppressive agents, affecting both innate and adaptive immunity (14, 15, 16). Recently, we have shown that VIP/PACAP affect bone marrow-derived DC (BM-DC) differently, depending on the DC maturation state (17). Immature DC treated with VIP or PACAP up-regulate CD86 expression, stimulate T cell proliferation and promote Th2-type responses while inhibiting the Th1-type proinflammatory response (17). In contrast, VIP and PACAP down-regulate CD80 and CD86 expression of LPS-matured DCs and inhibit their capacity to activate allogeneic or syngeneic T cells in vivo and in vitro (17). In the present study, we addressed the question whether VIP/PACAP induce tolerogenic BM-DCs. We report that VIP/PACAP induce CD11clowCD45Rbhigh DC that do not up-regulate CD40, CD80, or CD86, and secrete high amounts of IL-10 upon LPS stimulation. The VIP/PACAP-generated DCs induce Ag-specific Tr1-like Treg in vitro and in vivo, and these Treg are capable of transferring suppression to naive hosts.
Materials and Methods
Synthetic VIP, PACAP38, and PACAP6-38 were purchased from Calbiochem-Novabiochem. Capture and biotinylated Abs against murine IL-4, IL-5, IL-2, TNF-α, IL-6, IL-12, TGF-β1, and IFN-γ, mAbs to CTLA-4, CD40, CD4, Vβ3, CD11c, I-Ek, CD80, CD86, and CD103, and recombinant murine (rm)IFN-γ and rmIL-4 were purchased from BD Pharmingen. Capture and biotinylated Abs against murine IL-10 and rmGM-CSF and rmIL-2 were purchased from PreproTech. Rat anti-glucocorticoid-induced TNFR (GITR) was purchased from R&D Systems. Pigeon cytochrome c fragment (PCCF) was synthesized and purified by Research Genetics. H89 was obtained from ICN Pharmaceuticals. LPS (from Escherichia coli 055:B5), OVA, avidin-peroxidase, PMA, calphostin C, ionomycin, and monensin were purchased from Sigma-Aldrich. The VPAC1-agonist [K15,R16,L27]VIP1–7-GRF8–27 and the VPAC1-antagonist [Ac-His1,d-Phe2,K15,R16,L27] VIP3–7-GRF8–27 were previously described (18, 19).
B10.A (I-Ek), C57BL/6 (H-2b), and TCR-Cyt-5CC7-I/Rag1 transgenic (Tg, I-Ek) mice were obtained from The Jackson Laboratory and Taconic Farms. All mice used were between 7 and 12 wk of age. All animal procedures were approved by the Rutgers University Institutional Animal Care and Use Committee, on accordance with American Association for the Accreditation of Laboratory Animal Care guidelines.
Cell isolation and cultures
BM-DCs were generated from B10.A mice with the exception of the delayed-type hypersensitivity (DTH) experiments where DC were generated from C57BL/6 mice as previously described (17) in medium containing 200 U/ml (20 ng/ml) rmGM-CSF (5 × 106 U/mg) in the presence or absence of VIP or PACAP38 (from 10−12 to 10−6 M). Alternatively, VIP or PACAP (10−8 M) were added at different times after the initiation of culture. On day 6 or 7, nonadherent cells were collected by gently pipetting and centrifuged at 1200 rpm for 5 min. Cells were directly analyzed by flow cytometry or were seeded in flat-bottom 48-well microtiter plates (Corning Glass at 5 × 105 cells per well in a final volume of 500 μl and matured for 48 h with LPS (1 μg/ml) or CD40L-transfected Chinese hamster ovary cells (1 × 105 cells). In some experiments CD11c+ DCs were purified by immunomagnetic methods using anti-CD11c-conjugated magnetic beads and the AutoMACS system (Miltenyi Biotec).
Purified CD4 T cells from Tg mice were isolated by positive immunomagnetic selection with anti-CD4-conjugated magnetic beads (Miltenyi Biotec). The purified T cells were >98% CD4+ by FACS analysis. Effector Th1 and Th2 cells were generated as previously described (17). Briefly, purified naive CD4+ T cells (3 × 105 cells/ml) from Tg-PCCF mice were cultured with APCs (105 cells/ml) and PCCF (5 μg/ml) plus IL-2 (50 U/ml). Th2 polarization was performed in the presence of IL-4 (200 U/ml) plus anti-IFN-γ Ab (10 μg/ml), and Th1 polarization was performed in the presence of IFN-γ (1000 U/ml) and anti-IL-4 Abs (10 μg/ml). After 4 days, Th1 and Th2 effectors were characterized by intracellular cytokine profile (IFN-γ and IL-4, respectively).
APCs were prepared by T cell depletion of B10.A (I-Ek) spleen cells with a mixture of anti-CD8- and anti-CD4-conjugated magnetic beads and treated with 50 μg/ml mitomycin C (Sigma-Aldrich) for 20 min at 37°C.
DCs were also purified from spleen and lymph nodes as previously described with some modifications (12). In brief, spleen and mesenteric lymph nodes were cut into small fragments and digested with collagenase D (1 mg/ml). Resulting cells were layered over a Percoll gradient, and T and B cells were depleted by treating the recovered low-density cells with a mixture of mAbs (anti-CD3 and anti-B220) coupled to magnetic beads.
BM-DCs (1 × 106 cells/ml) were incubated with various mAbs (FITC-anti-CD80, FITC-anti-CD86, FITC-anti-CD40, FITC-anti-I-Ek, FITC-anti-CD45RB, PE-anti-CD11c, 2.5 μg/ml final concentration) at 4°C for 1 h. Isotype-matched Abs were used as controls, and IgG block (Sigma-Aldrich) was used to block nonspecific binding to Fc receptors. After extensive washing, the cells were fixed in 1% paraformaldehyde. Stained DC, gated according to forward- and side-scatter characteristics, were analyzed on a FACSCalibur flow cytometer (BD Biosciences). Fluorescence data were expressed as mean channel fluorescence (MCF), and as percentage (%) of positive cells after subtraction of background isotype-matched values.
For analysis of intracellular CTLA-4, T cells (1 × 106 cells/ml) were stained with PerCP-anti-CD4 mAb for 30 min at 4°C, fixed with Cytofix/Cytoperm solution (BD Pharmingen) and incubated for 45 min at 4°C with PE-anti-CTLA-4 mAb diluted in PBS + 1% BSA + 0.5% saponin. After extensive washing, cells were analyzed on a FACSCalibur flow cytometer. Surface GITR and CD103 expression was determined by FACS using PE-anti-CD103 and PE-anti-GITR Abs.
Mannose receptor-mediated endocytosis was measured as the cellular uptake of FITC-dextran (Sigma-Aldrich) and the fluid phase endocytosis through membrane ruffling was measured as the cellular uptake of Lucifer Yellow dipotassium salt (Sigma-Aldrich), and both were quantified by flow cytometry. Briefly, DCs (2 × 105 cells/sample) were incubated in medium containing FITC-dextran (1 mg/ml; m.w. 40,000) or with Lucifer Yellow (1 mg/ml) for 0, 60, and 120 min. Afterward incubation cells were washed twice in wash buffer, fixed in cold 1% paraformaldehyde, and analyzed by flow cytometry.
Assay of DC costimulatory activity
The costimulatory activity for syngeneic Ag-specific CD4 T cells was performed by using a TCR Tg model. In brief, various numbers of B10.A BM-DCs differentiated in the absence or presence of VIP/PACAP were added to purified PCCF-specific Tg CD4 T cells (5 × 105 cells/well) in the presence of PCCF (5 μM). Proliferation was determined by measuring BrdU incorporation as recommended by the manufacturer (Roche Applied Science). Cells cultured with an irrelevant Ag (OVA, 10 μg/ml) were used as control.
T cell anergy was determined after a secondary restimulation of T cells with mature DCs. In brief, B10.A BM-DCs (1 × 105 cells/well) differentiated in the absence or presence of VIP/PACAP were added to purified syngeneic PCCF-specific Tg CD4 T cells (5 × 105 cells/well) in the presence of PCCF (5 μM). After 3 days of culture, T cells were recovered by Ficoll gradient and DC depletion with anti-CD11c microbeads. T cells were rested for 2 to 4 days in complete medium supplemented with 2 U/ml IL-2, and restimulated (5 × 105 cells/well) with different numbers of LPS-matured B10.A DCs, generated in the absence of VIP/PACAP, and pulsed with PCCF (5 μM). Proliferation was determined by measuring the BrdU incorporation. Cell-free culture supernatants were harvested and used for cytokine determination by ELISA.
The cytokine content in DCs or DCs-CD4 T cell cultures was determined by sandwich ELISAs. The Ab pairs used were as follows, listed by capture/biotinylated detection Abs (BD Pharmingen): IL-4, BVD4–1D11/BVD6–24G2; IFN-γ, R4–6A2/XMG1.2; IL-5, TRFK5/TRFK4; IL-2, JES6–1A12/JES6–5H4; IL-12p40, C15.6/C17.8; TNF, MP6-XT22/MP6-XT3; IL6, MP5–20F3/MP5–32C11; IL-10, JES5–2A5/JES5–16E3.
For the intracellular cytokine analysis of restimulated CD4 T cells, 106 cells/ml were collected and stimulated with PMA (1 ng/ml) plus ionomycin (20 ng/ml) for 8 h. Monensin (1.33 μmol/ml) was added for the last 4 h of culture. Cells were stained with PerCP-anti-CD4 mAbs for 30 min at 4°C, washed, fixed/saponin permeabilized with Cytofix/Cytoperm, and stained with 0.5 μg/sample of FITC- and PE-conjugated anti-IFN-γ-, anti-IL-4-, or anti-IL-10-specific mAbs for 45 min at 4°C. Cells were analyzed by flow cytometry, using FITC/PE-conjugated isotypic mAbs as controls.
Total RNA was isolated from CD4 T cells or sorted CD4+CD25+ (106 cells) using the Ultraspec RNA reagent (Biotecx). Two micrograms of total RNA was reverse transcribed. Quantitative real-time RT-PCR was performed in an ABI PRISM cycler (Applied Biosystems) using a SYBR Green PCR kit from Applied Biosystems. A threshold was set in the linear part of the amplification curve, and the number of cycles needed to reach the threshold was calculated for every gene. Relative mRNA levels were determined by using standard curves for each individual gene and further normalization to hypoxanthine phosphoribosyltransferase (HPRT). Melting curves established the purity of the amplified band. Primer sequences are: neuropilin 1 (Nrp1) (5′-GCCTGCTTTCTTCTCTTGGTTTCA-3′, 5′-GCTCTGGGCACTGGGCTACA-3′); Foxp3 (5′-CTGGCGAAGGGCTCGGTAGTCCT-3′, 5′-CTCCCAGAGCCCATGGCA GAAGT-3′); HPRT (5′-TGGAAAGAATGTCTTGATTGTTGAA-3′, 5′-AGCTTG CAACCTTAACCATTTTG-3′).
In vitro T regulatory activity
B10.A BM-DCs differentiated in the absence or presence of VIP or PACAP were extensively washed, cultured with LPS (1 μg/ml) for 48 h at 37°C to induce DC activation/maturation, and added (5 × 104 cells/well) to purified syngeneic PCCF-specific Tg CD4 T cells (5 × 105 cells/well) in the presence of PCCF (5 μM). One week later, T cells were separated by Ficoll-Hypaque gradient, and DCs were removed by anti-CD11c microbeads (Miltenyi Biotec). Different numbers of purified T cells (Treg) were incubated with syngeneic PCCF-specific Tg CD4 T cells (5 × 105 cells/well) (responder T) in the presence of mitomycin C-treated B10.A spleen APCs (2 × 104 cells/well) and PCCF. The proliferation of responder CD4 T cells was determined by BrdU incorporation. In some experiments, cocultures were performed in the presence of blocking mouse anti-IL-10 (10 μg/ml), anti-TGF-β1 (40 μg/ml) and/or anti-CTLA4 (10 μg/ml) mAbs, or of rmIL-2 (100 U/ml). The anti-IL-10 and anti-TGF-β1 Abs were previously titrated. Concentrations leading to undetectable levels of IL-10 or TGF-β1 following Ab addition were used in the blocking experiments. Culture supernatants were assayed for IL-2 production by sandwich ELISA (BD Pharmingen).
Transwell experiments were done in 24-well plates (Millicell, 0.4 μm; Millipore). Responder PCCF-specific Tg CD4 T cells (5 × 105 cells) together with mitomycin C-treated B10.A spleen APCs (2 × 104 cells) and PCCF (5 μM) were placed in the bottom wells. T cells (5 × 105 cells) generated in the presence of VIP-DCs were placed together with mitomycin C-treated B10.A spleen APCs (2 × 104 cells) and PCCF (5 μM) in the upper wells. Three days later, the basket was removed, and the proliferation of the responder T cells was measured. IL-2 levels in culture supernatants were determined by ELISA.
Determination of Ab responses
Specific Ab responses in PCCF or OVA-immunized mice were determined by ELISA. Serum was obtained by cardiac puncture. Maxisorb plates (Millipore) were coated overnight at 4°C with 100 μl of soluble PCCF or OVA (10 μg/ml). The plates were incubated with serial dilutions of serum for 2 h at 37°C. Biotinylated anti-IgG, anti-IgG1, or anti-IgG2a (2.5 μg/ml) (Serotec) were added for 1 h at 37°C, followed by streptavidin-HRP. The bound enzyme was detected with the TMB substrate, and quantitated at OD450 in an ELISA reader (Bio-Tek Intrument). A standard curve was constructed for each assay by coating wells with an isotype-specific anti-mouse Ig followed by addition of known concentrations of the corresponding mouse Ig isotype.
Control- or VIP-DC were generated from C57BL/6 mice, pulsed with OVA (5 μM) for 2 h, and cultured with LPS (1 μg/ml) for 24 h to induce DC activation/maturation. To trace the injected cells in vivo, control-DC and VIP-DC were labeled with 1 ml of PBS, 0.1% BSA, 5 μM CFSE (Molecular Probes) for 10 min at 37°C, and washed three times in RPMI 1640 and 10% FBS. Control-DC, VIP-DC (3 × 105 cells) or saline (none) were injected s.c. into C57BL/6 mice previously immunized with OVA/CFA (100 μg, s.c.) (groups of nine). Four days after DC transfer, the mice were challenged with 50 μg OVA in 20 μl of PBS in the left ear with 20 μl of PBS in the right ear. The presence of the transferred CFSE-stained DCs in the draining lymph node (DLN) were determined by flow cytometry. Ear thickness was measured 24 h after challenge, and the thickness of the ear exposed to PBS was subtracted from the thickness of the ear exposed to OVA.
Determination of NF-κBp65 nuclear translocation and IkB phosphorylation
Nuclear and cytoplasmic protein extracts were isolated from DCs as previously described (20), and the content of NF-κBp65 and phosphorylated-IkB were determined by using a specific ELISA kit (Imgenex).
Characterization of VIP/PACAP-DC
To determine whether exposure to VIP/PACAP during DC differentiation results in DC phenotypical and functional changes, we compared DC generated from B10.A mice in the presence or absence of VIP/PACAP (VIP/PACAP-DC) in terms of surface markers, endocytic capacity, and cytokine production. Bone marrow cell cultures were initiated in medium containing GM-CSF and VIP or PACAP. We harvested the nonadherent cells 6 days later and determined the levels of surface MHC class II, CD40, CD80, and CD86 expression in the CD11c+ cell population (immature DC (iDC)). The DCs generated in the presence of VIP/PACAP express lower levels of CD11c, CD40, CD80, and CD86, and the reduction in costimulatory molecule expression by VIP/PACAP is dose-dependent (Fig. 1,A). Expression of MHC class II and of the costimulatory molecules CD40, CD80, and CD86 increases upon LPS stimulation of DC (mature DC (mDC)). However, similar to DC generated in the presence of 1,25(OH)2D3 or in the presence of TNF and IL-10, the VIP/PACAP-DC are resistant to LPS stimulation in terms of expression of CD40, CD80, and CD86 (Fig. 1 B).
To establish the requirements for exposure to VIP during DC differentiation, we compared the effects of VIP added at different times during DC generation. VIP added at day 0 or 1 results in reduction in CD40 and CD80 expression, and VIP added as late as day 3 still affects CD86 (Fig. 1C). In contrast, later additions (day 5 or 7) do not affect CD40 and CD80, while inducing CD86 in iDC, and down-regulate CD80 and CD86 in LPS-mDC (Fig. 1 C). These results are in agreement with the previously reported effects of VIP/PACAP added at the end of DC differentiation (day 7) (17).
Because iDC and mDC differ in their phagocytic capacity, we assessed the endocytic ability of VIP/PACAP-DC by using FITC-dextran or Lucifer Yellow. DC generated in the presence of VIP/PACAP exhibit a higher endocytic activity compared with control DC (Fig. 2,A). Upon TLR receptor activation, iDC mature into cells capable of producing high levels of TNF, IL-12, and IL-6. We determined the cytokine profile for VIP/PACAP-DC stimulated with LPS. In contrast to control DC which produce TNF, IL-12p40, and IL-6, and low levels of IL-10, the VIP/PACAP-DC produce extremely low levels of proinflammatory cytokines (TNF, IL-12, and IL-6), and high levels of IL-10 (Fig. 2,B). IL-10 producing tolerogenic DC, generated in the presence of TNF and IL-10 and expressing low levels of costimulatory molecules, have been previously described as being CD11clowCD45RBhigh. The VIP/PACAP-DC consist of cells with a similar phenotype, i.e., CD11clowCD45RBhigh (Fig. 2 C).
Taken together, these results indicate that the DCs generated in the presence of the neuropeptides differ from those generated with GM-CSF alone, in terms of costimulatory molecule expression, endocytic capacity, cytokine profile, and CD11cCD45RB phenotype. The characteristics expressed by the VIP/PACAP-DC are quite similar to those reported for tolerogenic DC generated in the presence of TNF and IL-10 (12).
VIP/PACAP-DC induce IL-10/TGF-β-producing T cells with low proliferative capacity
Tolerogenic DC are poor stimulators of T cell proliferation and cytokine production. To examine the capacity of the VIP/PACAP-DC to stimulate T cells, we treated the VIP/PACAP-DC with LPS or CD40L-transfected cells and cocultured them with syngeneic PCCF-specific T cells (Tg PCCF-T) in the presence of PCCF. The T cells exposed to the VIP/PACAP-DC did not proliferate (Fig. 2,D). Because LPS-stimulated VIP-DCs secrete large amounts of IL-10 (Fig. 2 B), we assessed the role of IL-10 in the induction of low proliferative T cells by adding anti-IL-10 Abs to the VIP-DCs/T cell cultures. The Ab dose necessary to neutralize the secreted IL-10 from LPS-stimulated VIP-DCs was determined in preliminary experiments. The neutralizing anti-IL-10 Abs did not affect T cell proliferation induced by control DCs (2.32 ± 0.31 in the absence of anti-IL-10 Abs and 2.37 ± 0.22 in the presence of anti-IL-10 Abs), but partially reversed the inhibitory effect of VIP-DCs on T cell proliferation (from 0.51 ± 0.12 in the absence of anti-IL-10 to 0.97 ± 0.08 in the presence of anti-IL-10 Abs).
To determine whether the T cells exposed to VIP/PACAP-DC are indeed anergic, we exposed transgenic PCCF-specific T cells (Tg PCCF-T) to control or VIP/PACAP-DC in the presence of PCCF, followed by re-exposure to fresh LPS-stimulated DC pulsed with PCCF. T cells exposed initially to control DC proliferated and produced IFN-γ, whereas the T cells exposed to VIP/PACAP-DC did not proliferate and did not produce IFN-γ upon re-exposure to PCCF-loaded DC (Fig. 2 E). This suggests that VIP/PACAP-DC induce anergic T cells and/or Treg.
Although neither type of T cells proliferate in vitro in response to Ag, Treg can release anti-inflammatory cytokines such as IL-10 and TGF-β. Therefore, we assessed the cytokine profile of T cells cocultured with VIP/PACAP-DC. VIP/PACAP-DC were stimulated with LPS and added together with PCCF to Tg PCCF-T cells. The T cells were purified 1 week later, restimulated with mitomycin C-treated splenic APC plus PCCF, and the cytokine profile was determined by ELISA. Control DC induced a predominant Th1 cytokine profile, with high levels of IFN-γ and IL-2. In contrast, T cells exposed to VIP/PACAP-DC produced IL-10, low levels of TGF-β, little IFN-γ, and no IL-2 (Fig. 3,A). T cells exposed repeatedly to LPS-stimulated VIP/PACAP-DC and PCCF (three weekly rounds), followed by purification and restimulation with PCCF/splenic APC, produced significantly higher levels of IL-10 and TGF-β, but no IFN-γ, IL-2, IL-5, and very little if any IL-4 (Fig. 3,B). Intracellular cytokine staining confirmed the high expression of IL-10 and to a much lesser degree of IL-4, and the reduction in IFN-γ in T cells exposed to VIP/PACAP-DC (Fig. 3 C).
Induction of tolerogenic DC by VIP is mediated through VPAC1 receptors and protein kinase A (PKA)
The biological effects of VIP/PACAP are mediated through three types of receptors, i.e., VPAC1, VPAC2, and PAC1 (21). We have previously reported that BM-DCs express VPAC1 and VPAC2 (17). To determine the VIP receptor involved in the generation of tolerogenic DCs, we used a VPAC1 agonist and VPAC1 and VPAC2/PAC1 antagonists. DC generated in the presence of the VPAC1 agonist exhibit a phenotype similar to VIP-DCs (CD11clowCD45RBhighCD40lowCD80lowCD86low) (Fig. 4,A). The role of VPAC1 was confirmed by the fact that the VPAC1 antagonist, but not the VPAC2/PAC1 antagonist, reversed the effect of VIP (Fig. 4,A). Because VPAC1 activates adenylate cyclase inducing intracellular cAMP, we assessed the role of PKA, one of the major cAMP targets. The PKA inhibitor H89, but not the protein kinase C inhibitor calphostin C, reversed the effect of VIP (Fig. 4 A).
The involvement of VPAC1 and PKA were further confirmed in experiments assessing the tolerogenic capacity of VIP-DCs. DCs generated in various conditions were stimulated with LPS for 24 h, followed by coculture with Tg PCCF T cells and PCCF. The purified T cells were restimulated 7 days later with mitomycin C-treated splenic APCs and PCCF and tested for proliferation and cytokine profile. T cells cultured with VIP-DCs or VPAC1-DCs did not proliferate and produced very low levels of IFN-γ, no IL-2, and high levels of IL-10 and TGF-β (Fig. 4, B and C). The effect of VIP was reversed by the VPAC1 antagonist and H89, but not by the VPAC2/PAC1 antagonist or calphostin C (Fig. 4, B and C).
LPS-induced NF-κB activation is prevented in VIP-DCs
Recently, inhibition of NF-κB has been reported to result in the induction of tolerogenic DCs (22). Therefore, we determined the effects of VIP on NF-κBp65 nuclear translocation and on IκB phosphorylation. VIP-DCs were stimulated with LPS, and the levels of NF-κBp65 and P-IκB were determined in nuclear and cytoplasmic extracts. In the absence of LPS stimulation, p65 was localized in the cytoplasm, and IκB was not phosphorylated (Fig. 4,D). Following LPS stimulation, p65 translocated to the nucleus in control DCs but not in VIP-DCs, and IκB was phosphorylated in control but not VIP-DCs (Fig. 4 D). The effect of VIP was reversed by the VPAC1 antagonist and by H89, suggesting that the VIP-induced inhibition of NF-κB activation is mediated by VPAC1 and PKA.
VIP/PACAP-DC induce functional Treg in vitro
When stimulated through the TCR, Treg suppress the proliferation and IL-2 production of Ag-specific effector T cells. To determine whether T cells exposed to VIP/PACAP-DC become functional Treg, we cultured Tg-PCCF CD4 T cells for 7 days with PCCF-pulsed DC generated from B10.A mice either in the presence or absence of VIP and matured with LPS. Tr(control) from cultures containing DC generated in the absence of the neuropeptide, and Tr(VIP) from cultures containing DC generated in the presence of VIP were purified and cocultured with fresh Tg-PCCF CD4 T cells in the presence of splenic APC pulsed with PCCF. Tr(VIP) inhibited the proliferation of fresh Tg-PCCF T cells, whereas the Tr(control) did not (Fig. 5,A). Similar results were obtained in respect to IL-2 production (Fig. 5,A, bottom panel). The inhibition of Tg-PCCF T cell proliferation by Tr(VIP) is dose-dependent (Fig. 5 A, central panel).
Next we addressed the question whether Tr(VIP) inhibit effector T cells through direct cellular contact and/or soluble factors. When Tr(VIP) and effector CD4 T cells are separated in Transwell experiments, the proliferation of effector T cells is still inhibited, but to a lesser degree, indicating that both direct contact and soluble factors mediate the inhibitory effect. In regular cocultures, addition of saturating amounts of anti-TGF-β, anti-IL-10, or anti-CTLA-4 Abs reversed inhibition only modestly. However, addition of both anti-IL-10 and anti-TGF-β has a more pronounced effect, and addition of all three Abs (anti-IL-10, anti-TGF-β, anti-CTLA-4) reverses the inhibitory effect completely (Fig. 5 B).
Because Treg were reported to express high levels of CTLA-4, and our experiments suggest that CTLA-4 plays a role in the inhibitory activity of Tr(VIP), we measured the levels of CTLA-4 in Tr(VIP) and Tr(control). Tr(VIP) express higher levels of CTLA-4, and the percentage of intracellular IL-10+CTLA-4+ cells is significantly higher in the Tr(VIP) population (Fig. 5,C). CD4+CD25+ Treg have been reported to express markers such as CD103, GITR, Nrp1, and Foxp3, although different Treg populations vary in the expression of these markers. Tr(VIP) and Tr(control) were restimulated with splenic APCs and PCCF, and stained with anti-CD4, anti-CD103 and anti-GITR Abs. In addition, sorted CD4+ T cells were analyzed for Foxp3 and Nrp1 expression by RT-PCR. Freshly isolated CD4+CD25+ nTregs expressing high levels of Foxp3, Nrp1, GITR, and CD103 were used as control. In contrast to nTreg, Tr(VIP) show little if any surface CD103 and GITR and no Nrp1 expression. There is a slight increase in Foxp3 expression in Tr(VIP) compared with Tr(control) but much less than in naturally occurring Treg (Fig. 5 C).
Tr(VIP) inhibit preferentially the response of Th1 effectors
Because VIP/PACAP were reported to favor Th2 over Th1 immunity in vivo (20), we compared the ability of Tr(VIP) to inhibit Th1 and Th2 effectors. Tr(VIP) and Tr(control) were cocultured with PCCF-specific Th1 or Th2 effectors, followed by measurements of proliferation and cytokine profile. Tr(VIP) inhibit Th1, and to a much lesser degree Th2 cell proliferation (Fig. 5,D), and reduce the production of Th1 cytokines (IFN-γ and IL-2), without significantly affecting the Th2 cytokines (IL-5 and IL-4) (Fig. 5 E).
VIP-DC induce Treg in vivo
DCs generated from B10.A mice in the presence or absence of VIP were stimulated with LPS, loaded with PCCF and injected into B10.A hosts that received 24 h previously Tg-PCCF CD4 T cells. Splenic CD4 T cells isolated 7 days later were restimulated in vitro with splenic APC/PCCF. T cells obtained from mice injected with control DC (generated without VIP) proliferate and secrete high levels of IFN-γ and IL-2 (Fig. 6,A). In contrast, T cells from mice injected with VIP-DC exhibit low proliferation, reduced levels of IFN-γ and IL-2, and high levels of IL-10 (Fig. 6,A). The suppressive activity of the T cells obtained from mice injected with VIP-DC was tested in cocultures with fresh Tg-PCCF CD4 T cells (responder T cells) in the presence of splenic APC/PCCF. In contrast to T cells from mice inoculated with control-DC, T cells from VIP-DC inoculated mice inhibit the proliferation of responder T cells (Fig. 6,B) and express higher levels of Foxp3 (Fig. 6 C). Therefore, the T cells developing in vivo in the presence of DC generated with VIP exhibit Treg characteristics and suppressive function, similar to the Tr(VIP) generated in vitro.
VIP-DC induce Ag-specific tolerance in vivo
DC generated from B10.A mice in the presence (VIP-DC) or absence of VIP (control-DC) were stimulated with LPS, loaded with PCCF, stained with CFSE, and injected i.v. into B10.A hosts that received 24 h previously Tg-PCCF CD4 T cells. The animals were immunized 1 week later with PCCF/CFA, followed by a boost 2 weeks later. The spleens were harvested 2 weeks after the second immunization. The presence of the inoculated DC was ascertained by analyzing spleen cells for CD11c and CFSE expression. We detected both control- and VIP-DC in spleen (Fig. 7 A).
Next, we assessed the proliferative capacity and the cytokine profile of splenic T cells following ex vivo restimulation with PCCF. Splenic T cells from mice that received control DC proliferate and produce high levels of IL-2 and IFN-γ and moderate amounts of IL-4 (Fig. 7, B and C). In contrast, splenic T cells from mice that received VIP-DC exhibit Treg characteristics such as reduced proliferation, reduced IL-2 and IFN-γ production, and high levels of IL-10 release (Fig. 7, B and C).
The mice injected with control-DC have both serum IgG1 and IgG2a anti-PCCF Abs, and in agreement with the previously described dominant inhibitory effect of Tr(VIP) on Th1 effectors, the anti-PCCF Abs in mice that received VIP-DC were almost exclusively of the IgG1 isotype (Fig. 7 D).
To determine whether the in vivo tolerance induced by VIP-DC is Ag-specific, we inoculated VIP-DC stimulated with LPS and loaded with PCCF in B10.A hosts, followed by immunization and T cell restimulation with OVA instead of PCCF. There were no differences in terms of proliferation and levels of anti-OVA IgG2a Abs between hosts that received VIP-DC (PCCF) or medium (Fig. 7, B and D). These results suggest that the tolerance induced by VIP-DC in vivo is restricted to the Ag presented by the inoculated DC.
VIP-DC inhibit DTH responses in vivo
Because VIP-DC seem to have a predominant effect on Th1 cells, we used an in vivo model of DTH. Control- and VIP-DC generated from C57BL/6 mice were pulsed with OVA, stimulated with LPS, stained with CFSE, and injected s.c. in OVA-immunized C57BL/6 mice. The mice were challenged with OVA in the left ear pinna 4 days later, and the DTH reaction was assessed by measuring ear thickness in comparison to the right ear control (PBS). The presence of control- and VIP-DC in the draining lymph nodes was assessed by analyzing the lymph node cells for CFSE and CD11c staining. CFSE-labeled CD11c+ cells were detected in the draining lymph nodes, with VIP-DC seemingly having undergone fewer divisions (Fig. 8 A).
Mice that received control-DC developed DTH reactions higher than controls (no DC), whereas those receiving VIP-DC exhibit reduced DTH reactivity (Fig. 8,B). In addition, we also assessed the proliferation of T cells isolated from the draining lymph nodes following ex vivo restimulation with OVA, and the levels of anti-OVA Abs in serum. T cells from mice inoculated with control-DC proliferate at higher levels than controls (no DC) and, again, inoculation of VIP-DC results in a substantial reduction in T cell proliferation for all OVA concentrations tested (Fig. 8,C). Similarly, mice inoculated with control-DC produce high levels of anti-OVA Abs, whereas those inoculated with VIP-DC have a level of anti-OVA Abs below the controls (no DC) (Fig. 8 D).
Transfer of tolerance with CD4 T cells from VIP-DC inoculated donors
Ag-specific tolerance can be adoptively transferred with Treg. We investigated the potential of Treg generated by VIP-DC to transfer tolerance to naive hosts. DCs generated from B10.A mice in the presence or absence of VIP were stimulated with LPS, loaded with PCCF, and injected i.v. into B10.A hosts that received 24 h previously Tg-PCCF CD4 T cells. Splenic CD4 T cells harvested 7 days later were transferred i.v. into naive B10.A mice, which were immunized a week later with PCCF/CFA and received a boost 2 weeks later. Splenic T cells were analyzed 2 weeks after the boost in terms of proliferation and cytokine profile following ex vivo restimulation, and serum anti-PCCF levels were determined. T cells from donors inoculated with VIP-DC are capable of transferring the suppressive activity into B10.A naive hosts, as determined by the reduction in proliferation, IL-2 and IFN-γ production, and increased IL-10 production by host splenic T cells (Fig. 9, A and B). In agreement with the substantial reduction in Th1 cytokines, the levels of anti-PCCF IgG2a Abs in hosts receiving T cells from the VIP-DC inoculated donors were significantly reduced (Fig. 9 C).
In vivo induction of tolerogenic DC by VIP
In the experiments described above, the VIP-induced tolerogenic DC were generated in vitro. Next, we addressed the question whether VIP can induce tolerogenic DC in vivo, by administering VIP together with PCCF (Ag) to Tg PCCF mice. Controls were injected with Ag alone. Eight days later, spleen and mesenteric lymph node cells depleted of T and B cells consisted of approximately the same percentage of CD11c+ (DC) in both controls and VIP-injected mice (26–29%) (Fig. 10,A). However, the CD45RBhigh population increased from 7.2% (Ag control) to 17.3% (VIP+Ag), and these cells expressed lower levels of CD11c, similar to the profile of the VIP-DC generated in vitro (Fig. 2 C).
Also, similar to the in vitro generated VIP-DCs, the CD11c+ cells generated in vivo following VIP administration expressed lower levels of CD40, CD80, and CD86 following LPS stimulation (Fig. 10,B). To determine whether the VIP-DCs generated in vivo function as tolerogenic DCs, we stimulated T- and B-depleted spleen and lymph node cells with LPS, isolated the CD11c+ DCs and cultured them with purified Tg PCCF T cells in the presence of PCCF. Four days later, we measured proliferation and cytokine release. The cultures containing the in vivo generated VIP-DCs did not proliferate, produced much less IFN-γ and IL-2, increased levels of TGF-β, and significantly higher amounts of IL-10 (Fig. 10 C). This suggests that, similar to the in vitro VIP-DCs, the in vivo VIP-DCs induce IL-10/TGF-β-producing T cells with low proliferative capacity.
To verify that the IL-10/TGF-β producing T cells induced in cocultures with the in vivo generated VIP-DCs (Ag+VIP) function as Treg (Tr), we added different numbers of Tr cells to syngeneic PCCF Tg T cells (responder T cells) in the presence of splenic APCs and PCCF. The proliferation of the responder T cells was inhibited in a dose-dependent manner (Fig. 10 D). These results indicate that in vivo administration of VIP results in the induction of tolerogenic DCs, which in turn can induce IL-10/TGF-β-producing Treg.
VIP and PACAP are potent immunosuppressive agents that affect both innate and adaptive immunity (14, 15, 16). Until now, the mechanisms described for their immunosuppressive activity included macrophage/dendritic cell/microglia deactivation, and support for Th2 effector differentiation and survival. In this study, we report on the induction of Treg through the VIP/PACAP generation of tolerogenic DCs.
We reported previously that VIP/PACAP have different effects on immature and LPS-matured DCs (17). The VIP/PACAP treatment of immature DCs led to the up-regulation of CD86, and increased capacity to stimulate CD4 T cells and promote Th2-type responses. In contrast, the VIP/PACAP treatment of LPS-matured DCs prevented the expression of CD80 and CD86, reduced the stimulatory activity for CD4 T cell proliferation and cytokine production, and differentially affecting Th1- and Th-2 chemoattractants (23, 24). These results support the previously reported effect of VIP/PACAP on the immune response, i.e., the inhibition of the proinflammatory, Th1-dependent response, with a concomitant bias toward Th2 responses.
In this study, we generated BM-DCs in the presence of VIP/PACAP. The addition of the neuropeptides early during DC differentiation (up to day 3) results in DCs phenotypically resembling the Tr1-inducing DCs described by Wakkach et al. (12). Similar to DCs generated in the presence of GM-CSF, TNF, and IL-10 (12), the DCs generated with VIP/PACAP are CD11clowCD45Rbhigh, do not up-regulate the expression of CD80 and CD86, and secrete IL-10 but not IL-12, following LPS treatment. Similar CD11clowCD45Rbhigh tolerogenic DC were also generated in vivo upon administration of VIP and Ag. These results suggest that VIP and PACAP provide a maturation signal similar to IL-10 (12) or to the active form of vitamin D3 1,25(OH)2D3 (8, 9, 25, 26, 27) leading to the differentiation of tolerogenic “semimature” or “quiescent” DCs (28, 29, 30).
The effect of VIP in generating tolerogenic DCs is mediated through the VPAC1 receptor and the cAMP/PKA signaling pathway. VPAC1 has been previously identified as the major functional receptor in dendritic cells, macrophages, and bone marrow cells (14, 17, 31). Previously VIP has been reported to inhibit NF-κBp65 nuclear translocation, DNA binding and transactivating activity in macrophages (20). Here, we report that both NF-κBp65 nuclear translocation and IκB phosphorylation are inhibited in VIP-DCs. The connection between NF-κB transactivating activity, CD40 expression, and DC function has been established recently. The association between tolerogenic DCs and lack of CD40 expression or signaling has been demonstrated both in vivo and in vitro (32, 33). Expression of CD40 depends on NF-κB, primarily p65 (32, 33). A recent report by Yang et al. (22) demonstrates that inhibition of NF-κB in DCs leads to failure of CD40, CD80, and CD86 expression upon LPS stimulation, and to the generation of tolerogenic DCs. Therefore, we propose that the mechanism by which VIP/PACAP induce tolerogenic DCs involves the VPAC1 → cAMP → PKA-mediated inhibition of IκB phosphorylation and NF-κBp65 nuclear translocation, leading to lack of CD40 expression.
Dendritic cells play an important role in the differentiation of peripherally induced Tr1 and Th3/Tr2 Treg (5, 28, 36, 37). The generation of Treg by VIP-DCs is partially mediated through the release of VIP-DCs-derived IL-10, as shown by the effect of neutralizing anti-IL-10 Abs. However, because saturating concentrations of anti-IL-10 could not restore >50% of T cell proliferation, other VIP-DCs related factors must play a role. The best candidate is the reduced CD40 expression on VIP-DCs, because blockade of CD40 signaling has been shown to promote tolerance in immature DCs (38, 39, 40).
CD4+ T cells exposed to VIP/PACAP-generated DCs do not proliferate and secrete large amounts of IL-10, some TGF-β, and no IL-2 or IFN-γ, following re-exposure to stimulatory Ag-pulsed DCs. The cytokine profile resembles best the previously described Tr1 cells (12). In addition, CD4 T cells exposed to VIP/PACAP-generated DCs function as Treg, inhibiting the proliferation and IL-2 production of responder T cells. These cells express high levels of CTLA-4, but low levels of other Treg markers such as Foxp3, CD103, GITR, and neuropilin 1. The suppressive function appears to be mediated through both soluble factors (IL-10 and TGF-β) and direct cellular contact. CTLA-4 may play a role in suppression through direct cellular contact, because the combination of anti-IL-10, anti-TGF-β, and anti-CTLA-4 Abs completely reverses suppression. Although CTLA-4 is expressed at high levels in Treg, its role in the development and/or function of Treg is not clear. This is partly because CTLA-4-deficient mice have normal Treg development and homeostasis. However, these mice exhibit increased levels of IL-10 and TGF-β, which suggests the existence of compensatory mechanisms (41). In support for a CTLA-4 role in Treg suppression, CTLA-4 binding to B7 has been shown to induce IDO in DCs, leading to tryptophan depletion and induction of proapoptotic molecules (42).
CD4+CD25+ nTreg have been characterized by high expression of Foxp3 and GITR, and recently of neuropilin 1 (43, 44, 45, 46). In contrast, the Tr1 cells generated with VIP/PACAP-DCs express low levels of Foxp3, and very little, if any, Nrp1 and GITR. In agreement with our results, Tr1 cells generated by repetitive stimulation with IL-10-secreting iDCs have been shown to express low levels of CD25 and Foxp3 (7).
Although all DC subsets are generated from bone marrow progenitors, recent studies indicate that with the exception of plasmacytoid DCs, all the other DC subsets differentiate at their peripheral homing site (47). Blood monocytes differentiate into macrophages or DCs upon migration into the tissues, with cytokines and environmental factors playing an important role in shifting the balance between these cell types (48, 49). In physiological conditions, VIP could affect DCs differentiation in the bone marrow as well as in the periphery. Indeed, VIP-ergic nerve fibers have been identified in the bone marrow, as well as skin, gastrointestinal tract, and secondary lymphoid organs (50, 51). In addition, activated Th2 effectors have been shown to release functional VIP (52). Therefore, in steady-state conditions, VIP released from the innervation or secreted by immune cells could act as an endogenous maturation signal driving the differentiation of tolerogenic DCs loaded with self- or commonly encountered Ags.
In addition, the VIP-induced generation of tolerogenic DCs in vitro could lead to new therapeutic developments in autoimmune diseases and in allogeneic transplantation. We showed that in vitro pulsing of tolerogenic VIP-DCs with Ag followed by administration in vivo leads to Ag-specific tolerance mediated by Treg. We would like to propose that the in vitro generation of VIP-DCs followed by loading with specific Ags might provide a more targeted approach for the induction of endogenous Ag-specific Treg in the treatment of autoimmune diseases.
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 supported by National Institutes of Health Grants AI52306 and AI47325 (to D.G.), Johnson & Johnson Neuroimmunology Fellowships (to M.D.), and the Spanish Ministry of Health PI04/0674 (to M.D.).
Abbreviations used in this paper: DC, dendritic cell; BM-DC, bone marrow-derived DC; DLN, draining lymph nodes; iDC, immature DC; mDC, mature DC; DTH, delayed-type hypersensitivity; GITR, glucocorticoid-induced TNFR; MCF, mean channel fluorescence; Nrp1, neuropilin 1; PACAP, pituitary adenylate cyclase-activating polypeptide; PCCF, pigeon cytochrome c fragment; PKA, protein kinase A; Tg, transgenic; Treg, regulatory T cell; nTreg, naturally occurring CD4+CD25+ Treg; iTreg, induced CD4+CD25+ peripheral Treg; VIP, vasoactive intestinal peptide.