A. M., S. R. and V. H. equally contributed to this manuscript.
Cellular immune response
CpG ODN enhance antigen-specific NKT cell activation via plasmacytoid dendritic cells
Article first published online: 18 JUL 2005
Copyright © 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
European Journal of Immunology
Volume 35, Issue 8, pages 2347–2357, August 2005
How to Cite
Marschner, A., Rothenfusser, S., Hornung, V., Prell, D., Krug, A., Kerkmann, M., Wellisch, D., Poeck, H., Greinacher, A., Giese, T., Endres, S. and Hartmann, G. (2005), CpG ODN enhance antigen-specific NKT cell activation via plasmacytoid dendritic cells. Eur. J. Immunol., 35: 2347–2357. doi: 10.1002/eji.200425721
- Issue published online: 4 AUG 2005
- Article first published online: 18 JUL 2005
- Manuscript Accepted: 21 JUN 2005
- Manuscript Revised: 4 MAY 2005
- Manuscript Received: 5 OCT 2004
- CpG oligonucleotide;
- NKT cells;
- Plasmacytoid dendritic cell;
- Myeloid dendritic cell;
- Type I interferon
Human Vα24+ Vβ11+ natural killer T cells (NKT cells) are “natural memory” T cells that detect glycolipid antigens such as α-galactosylceramide (α-GalCer) presented on CD1d. In the present study we found that highly purified Vα24+ NKT cells lack TLR9 mRNA, and thus are not sensitive towards stimulation with CpG oligodeoxynucleotides (ODN). Within PBMC, however, CpG ODN synergistically activated NKT cells stimulated with their cognate antigen α-GalCer. Depletion of plasmacytoid dendritic cells (PDC) or myeloid dendritic cells (MDC) revealed that both DC subsets were necessary for the synergistic activation of NKT cells by α-GalCer and CpG ODN. While PDC were responsible for the stimulation of NKT cells with CpG ODN, MDC but not PDC presented α-GalCer via CD1d. Partial activation of NKT cells was mediated by PDC-derived IFN-α, whereas full activation of NKT cells as indicated by IFN−γ production required cell-to-cell contact of PDC and NKT cells in addition to IFN-α; OX40 was involved in this interaction. We conclude that CpG-activated PDC enhance α-GalCer-specific NKT cell activation, and bias activated NKT cells towards a Th1 phenotype. Our results lead to a novel concept of PDC function: to regulate effector activity of antigen-stimulated T cells in a cell contact-dependent manner without the need of simultaneous presentation of the cognate T cell antigen.
Natural killer T (NKT) cells represent a small subset of non-conventional innate T cells with a restricted TCR repertoire for the recognition of glycolipid antigens presented on CD1d, an MHC class I-like molecule. NKT cells express markers of NK cells and exhibit an activated memory phenotype. Depending on the microenvironment and the type of stimulation, they can release large amounts of both Th1 and Th2 cytokines (especially IL-4 and IFN-γ) and show cytotoxic activity in vitro (reviewed in 1–3). In both mice and man NKT cells express a homologous semi-invariant TCR that in humans consists of a Vα24 chain preferentially paired with a Vβ11 chain 4. In mice, NKT cells are most frequent in the liver (about 30% of hepatic T cells), bone marrow and thymus (reviewed in 5). Smaller NKT cell populations are present in spleen and blood (reviewed in 5). Similar to their murine counterparts, human NKT cells preferentially accumulate in the liver though with a far lower frequency than in mice (4% of hepatic T cells) (6, 7 and reviewed in 1, 5).
Due to the conserved TCR, both mouse and human NKT cells recognise the same specific glycolipid antigen α-galactosylceramide (α-GalCer) when presented by CD1d molecules on antigen-presenting cells (APC) 8–10. Originally isolated from a marine sponge, α-GalCer was first described as an agent with strong anti-metastatic activity in mice and later found to specifically activate NKT cells in a CD1d-restricted manner 8, 11, 12. Glycolipid antigens similar to α-GalCer have been detected in certain bacteria and under some abnormal conditions in mammalian tissue such as cancer cells, suggesting that α-GalCer recognition by NKT cells represents a conserved defence mechanism activated by special glycolipid-carrying pathogens and stress signals 8, 13.
In vivo injection of α-GalCer in mice induces a rapid cascade of cellular activation events, beginning with the activation of NKT cells and propagating to other cells, such as NK cells, B cells, T cells and dendritic cells (DC) 5, 14. Studies in mice have revealed that α-GalCer treatment promotes resistance against infections, tumours and autoimmunity 15–27. In a murine vaccination study, α-GalCer potentiated protective immune responses induced by malaria vaccines 28. Clinical trials using α-GalCer for the treatment of human cancer have been initiated 29. The striking conservation of α-GalCer recognition raises optimism that the results obtained by studies in mice can also be achieved in humans.
CpG oligodeoxynucleotides (CpG ODN) represent another group of compounds that have shown promise as vaccine adjuvant 30 and for immunotherapy of tumours 31. In the vertebrate immune system Toll-like receptor 9 (TLR9) has evolved to detect unmethylated CG dinucleotides with certain flanking bases (CpG motifs) in microbial DNA 32–34. TLR are pattern recognition receptors of the innate immune system for the detection of pathogen-derived microbial molecules 35. The plasmacytoid DC (PDC) is specialised in the detection of viruses; it employs TLR9 to detect CpG motifs in viral DNA and produces large amounts of type I IFN (IFN−α and IFN−β), the signature cytokines of an anti-viral immune response 36–39. CpG-activated PDC indirectly activate other immune cells such as monocytes 40, 41, NK cells, γ/δ T cells 42 and memory α/β T cells 43, 44. CpG ODN are potent humoral vaccine adjuvants in primates 45–49 and in humans 50.
In the present study we hypothesised that CpG ODN and the NKT cell antigen α−GalCer in combination may potentiate each other in a way that could be useful for an improved immunotherapy of infectious diseases and cancer. We demonstrate strong adjuvant activity of CpG ODN for α−GalCer-specific NKT cell responses. Our results reveal that PDC modulate the effector function of antigen-stimulated T cells by cell contact without presenting the cognate antigen.
CpG ODN activate NKT cells within PBMC and enhance α-GalCer-specific NKT cell responses in a Th1-biasing manner
We positively selected NKT cells from PBMC using a Vα24-specific antibody and magnetic beads (see the methods section). Approximately 40% of isolated Vα24+ cells (mean: 45 ± 12%; n=6) expressed the NKT cell-specific Vβ11 chain (Fig. 1). To study whether NKT cells are equipped to directly recognise CpG ODN, we examined TLR expression of purified NKT cells by quantitative real time RT-PCR. Vα24+ cells expressed TLR1 and TLR5 mRNA, whereas TLR9 mRNA was very low (6 transcripts per 103 copies of the housekeeping gene cyclophilin B) when compared to PDC (109 TLR9 transcripts) (Fig. 2). Consistent with the low level of TLR9, isolated NKT cells were not responsive to CpG ODN (data not shown).
In humans, TLR9-expressing PDC are sensitive to CpG ODN, whereas monocytes, myeloid DC (MDC), T cells and NK cells are indirectly activated via PDC-derived cytokines 42, 43, 51. To examine whether similar indirect effects of CpG ODN contribute to activation of human NKT cells, we analysed NKT cell activation by CpG ODN within PBMC. The mean frequency of NKT cells in PBMC was found to be 0.3% (± 0.04%; n=24) of all CD3+ lymphocytes. The analysis of the rare NKT cell population was facilitated by adding purified Vα24+ cells from the same donor to the PBMC tested (to a final concentration of 5–12% Va24+ cells in PBMC). In the following experiments, this cell preparation is referred to as ‘NKT cell-enriched PBMC'.
NKT cell-enriched PBMC were stimulated with CpG-A ODN 2216 52 in the presence or absence of the NKT cell-specific antigen α-GalCer. In the absence of antigen, CpG ODN induced significant up-regulation of CD69 on NKT cells (Fig. 3A, left panel). Together with α-GalCer, CpG ODN synergistically increased the antigen-dependent CD69 up-regulation. This effect was CpG specific as the GC control ODN of ODN 2216 did not increase CD69 expression on NKT cells compared to medium or α-GalCer alone (Fig. 3A, right panel).
To examine whether stimulation with CpG ODN influences the cytokine production of NKT cells, we analysed IFN-γ and IL-4 in the supernatants of NKT cell-enriched PBMC cultures. CpG ODN stimulation induced low but significant IFN-γ production in the absence of α-GalCer (medium: 40 ± 26 pg/mL; ODN 2216: 163 ± 58 pg/mL; n=19; Fig. 3B, left panel, data as fold increase compared to condition with α-GalCer). Like for CD69 expression, the α-GalCer-dependent IFN-γ production was synergistically increased after costimulation with CpG ODN (absolute values: α-GalCer: 2690 ± 1074 pg/mL; α-GalCer plus ODN 2216: 4244 ± 1370 pg/mL; n=19). In contrast, NKT cells produced negligible amounts of IL-4 upon stimulation with α-GalCer with and without CpG ODN (medium: 0.05 pg/mL; ODN 2216: 0.55 pg/mL; α-GalCer: 4.40 pg/mL; α-GalCer plus ODN 2216: 5.45 pg/mL; n=2; data not shown), indicating that CpG ODN leads to a Th1 bias of α-GalCer-stimulated NKT cells. To confirm that the cytokines detected in the supernatants were indeed derived from NKT cells, we compared PBMC and NKT cell-enriched PBMC of the same donor with regard to IFN-γ production after α-GalCer with or without CpG ODN stimulation. CpG ODN induced high levels of α-GalCer-dependent IFN-γ only in NKT cell-enriched PBMC (Fig. 3C).
Identification of MDC as the main APC for NKT cells within human PBMC
The glycolipid antigen α-GalCer is presented to NKT cells via the MHC class I-like molecule CD1d expressed on APC 9, 10. Because CpG ODN enhance α-GalCer-mediated stimulation, we asked whether PDC express CD1d and thus are able to present α-GalCer to NKT cells. Using flow cytometry, we found that PDC lack CD1d, while MDC expressed high levels of CD1d (Fig. 4A). The expression of CD1d on MDC was much higher than on PBMC (data not shown). Depletion of MDC markedly reduced IFN-γ production of NKT cell-enriched PBMC after α-GalCer stimulation (Fig. 4B). Thus, in peripheral blood CD1d-expressing MDC play an important role in NKT cell activation by α-GalCer.
PDC contribute to CpG ODN-induced NKT cell activation via IFN-α
We hypothesised that the CpG ODN-mediated enhancement of NKT cell activation is indirectly mediated via TLR9-expressing PDC. To test this hypothesis, PDC were depleted from PBMC, and the PDC-depleted PBMC were enriched with isolated NKT cells. As shown in Fig. 5A, PDC depletion abolished CpG ODN-induced α-GalCer-dependent IFN-γ production in NKT cells indicating that CpG-induced activation of NKT cells requires the presence of PDC.
PDC produce a variety of cytokines and chemokines upon stimulation with CpG ODN, including type I IFN (IFN-α and IFN-β) 39. We found that CpG ODN-conditioned PDC-derived medium was as effective as the addition of CpG ODN to enhance CD69 expression on NKT cells (Fig. 5B). Recombinant IFN-α also enhanced α-GalCer-dependent CD69 expression, but to a lower extent (Fig. 5B). In addition, blockade of type I IFN at the beginning of NKT cell activation within PBMC abolished CpG ODN-induced α-GalCer-dependent IFN-γ production (Fig. 6A). In contrast, blockade of IL-12 did not reduce CpG-mediated IFN-γ induction in NKT cells (Fig. 6B). Thus, PDC-derived IFN-α was required for synergistic activation of α-GalCer-stimulated NKT cells.
PDC-derived soluble factors, in turn, could affect NKT cells directly or via activation of accessory cells. Indeed, MDC stimulated with PBMC-derived CpG ODN-conditioned medium up-regulated not only CD80, CD86, and CD40 (CD86 and CD40 not in figure) but also the maturation marker CD83 (Fig. 7). Recombinant IFN-α induced similar changes, but to a lower extent. Thus, CpG-induced soluble factors and recombinant IFN-α enhance activation and maturation of MDC, supporting MDC-dependent NKT cell activation.
CpG ODN-triggered IFN-γ production in NKT cells requires cell-to-cell contact with PDC
In contrast to CD69 up-regulation (see Fig. 5B), unexpectedly, neither PDC-derived CpG ODN-conditioned medium nor recombinant IFN-α enhanced α-GalCer-dependent IFN-γ production in NKT cells (Fig. 8A). Thus, although type I IFN is required for CpG-mediated IFN-γ induction in NKT cells (see Fig. 6A), PDC-derived soluble factors or recombinant IFN-α are not sufficient to enhance antigen-dependent IFN-γ production. This suggested that at least for IFN-γ production in NKT cells also a cell contact-dependent mechanism of CpG-activated PDC might be relevant. To further investigate this, we used a transwell system to separate NKT cells (PDC-depleted NKT cell enriched PBMC) from PDC. In this setting, neither CpG ODN 2216 or CpG ODN 2006 53 enhanced α-GalCer-dependent IFN-γ production of NKT cells (Fig. 8B, left panel). In contrast, CpG-mediated α-GalCer-dependent CD69 up-regulation on NKT cells was still intact (Fig. 8B, right panel). Together these data indicate that, while PDC-derived soluble factors are sufficient for CD69 up-regulation on NKT cells, IFN-γ production in NKT cells requires both direct cell-to-cell contact with PDC and PDC-derived type I IFN.
The mechanism responsible for the cell-contact-dependent activation of NKT cells by CpG-stimulated PDC was studied using the model antigen anti-CD3 beads instead of the NKT cell-specific antigen α-GalCer. With anti-CD3, the presence of MDC is not required for NKT cell stimulation. In a co-culture of NKT cells and PDC, we examined the contribution of OX40L using blocking antibodies. We found that the blockade of OX40L markedly inhibited the ability of PDC to support antigen-induced IFN-γ production in NKT cells (Fig. 8C, left panel). In contrast, CD69 expression on NKT cells was not reduced. Next, we were interested whether CpG-stimulated PDC enhance antigen-dependent IFN-γ production not only in NKT cells but also in CD8 T cells. Since anti-CD3 activates the TCR of both NKT cells and CD8 T cells, the same experimental system could be used. Indeed, the presence of CpG-stimulated PDC dramatically increased the ability of CD8 T cells to produce IFN-γ upon antigen stimulation, and, similar to NKT cells, this effect could be largely inhibited by adding increasing concentrations of blocking antibodies against OX40L. Together these results indicated that OX40L on activated PDC represents a major mechanism by which CpG-stimulated PDC support IFN-γ synthesis of NKT cells, and that this activity of PDC is not restricted to NKT cells but is also effective in enhancing IFN-γ production in CD8 T cells.
T cells are thought to integrate two signals when in contact with APC: first, the recognition of their cognate antigen, and second, costimulatory molecules expressed on the surface of, or released by, the cell presenting the cognate antigen. Here we provide evidence that, in addition to antigen (first signal) and costimulatory molecules on the APC (second signal), appropriately activated PDC provide a third signal. By analysing the interaction of NKT cells with MDC and PDC, we were able to separate antigen presentation from PDC-derived signals: MDC but not PDC expressed the antigen presenting molecule CD1d (for NKT cells), and PDC but not MDC carry the CpG-receptor TLR9.
We demonstrate that, in a co-culture of NKT cells, MDC and PDC, CpG synergistically increased the expression of CD69 and the production of IFN−γ in antigen (α-GalCer)-stimulated NKT cells. The induction of IFN−γ but not IL-4 indicated that appropriately stimulated PDC bias α-GalCer-activated NKT cells towards Th1 and not to the NKT0 phenotype that has been described for NKT cells in the presence of unstimulated PDC 54. It is interesting to note that, in this setting, the classic Th1-inducing cytokine IL-12 55 was not involved in IFN−γ induction.
No direct effect of CpG ODN on purified NKT cells was observed. This was in agreement with the lack of TLR9 in purified NKT cells. Similar to conventional αβ T cells and NK cells 43, the only two TLR expressed at considerable levels in NKT cells were TLR1 and TLR5. Like NKT cells, MDC were unable to respond to CpG ODN directly. However, indirect activation and differentiation of MDC (CD80, CD83) via PDC-derived cytokines contributed to partial activation of NKT cells. Of note, PDC-derived cytokines did not up-regulate the expression of CD1d on MDC or PDC (data not shown).
Full activation of NKT cells as evidenced by the production of large amounts of IFN−γ required both PDC-derived IFN−α and cell-to-cell contact between PDC and NKT cells. Only partial activation of NKT cells (CD69) was observed with PDC-derived IFN−α alone. Partial activation of NKT cells by PDC-derived IFN−α is in agreement with our previous studies on NK cells and γδ T cells 42 and on memory T cells 43. However, only the use of NKT cells in this study allowed us to analyse cell-contact dependent antigen-specific T cell activation by PDC with presentation of the cognate antigen on a separate cell. In contrast, γδ T cells detect antigen adsorbed to any cell surface without the need of a presenting molecule, and MHC class I and II molecules for antigen presentation to memory T cells are expressed on both PDC and MDC.
The role of PDC in virus recognition and type I IFN production is well established. However, there is an ongoing debate whether PDC fulfill the classic function of a DC, which is the initiation of primary T cell responses 56. It has been demonstrated that PDC fail to induce proliferation of naive T cells, but promote expansion and Th1 differentiation of antigen-experienced unpolarised T cells 57. Others have reported that PDC directly primed naive CD8 T cells during viral infection 58. Furthermore, PDC were found to prime antigen-specific CD8 T cell responses in vitro59. Our own studies showed that PDC were required for enhanced generation of cytotoxic T cells in PBMC with CpG ODN, but that purified PDC were weak at priming of CD8 T cell responses 44. Based on our results, we propose a model in which PDC contribute to the effector function of NKT cells without the need of simultaneous presentation of the cognate antigen by the PDC. Consistent with our model of a TCR-independent interaction of PDC with NKT cells, Kadowaki and colleagues 54 observed that anti-CD1d antibodies blocked the interaction between NKT cells and monocyte-derived DC, but had no effect on the interaction of NKT cells with PDC.
It has been reported that PDC express OX40L upon stimulation, and that virus-stimulated PDC modulate Th responses via OX40L and IFN-α 60. However, involvement of OX40L in PDC-mediated activation of NKT cells or CD8 T cells has not been studied. In our studies we demonstrate that OX40L contributes to the enhancement of antigen-dependent IFN-γ production by CpG-stimulated PDC both in NKT cells and CD8 T cells. Our data extend earlier studies in the literature in which OX40-mediated costimulation was found to regulate the accumulation of antigen-stimulated CD8 cells 61. Of note, in our studies the Th1 cytokine IFN-α was required (but not sufficient) for the PDC-mediated enhanced IFN-γ production in NKT cells. Partial but not complete inhibition of PDC-mediated enhanced IFN-γ production by blockade of OX40L function indicates that other costimulatory molecules in addition to OX40L might be involved in the PDC-NKT cell interaction leading to a Th1 bias of NKT cell cytokine production. Furthermore, OX40L is not selectively expressed on PDC, and consequently, OX40L on MDC may also contribute to NKT cell activation in mixed cell cultures or in vivo.
Our results have implications for the use of α-GalCer for immunotherapy of infectious diseases and cancer. There are a number of uncertainties with regard to therapeutic targeting of NKT cells. The proposed applications of NKT cell activation range from suppression of autoimmunity to tumour rejection 3. NKT cell differentiation towards a Th1 or Th2 direction was often unpredictable 62. NKT cells are less frequent in humans than they are in mice 2, and high doses of α-GalCer have caused severe liver damage in mice 63. Furthermore, there is evidence that in the absence of cognate antigen, NKT cells restrain the anti-tumour activity of CpG ODN in mice 64. The combination of CpG ODN and NKT cell antigens such as α-GalCer may have a number of advantages: (1) NKT cells will be biased towards Th1 and turned away from their immunosuppressive state; (2) NKT cell effects will be enhanced even for relatively small numbers of NKT cells as present in humans; (3) higher activity will allow the dose of α-GalCer to be reduced to avoid NKT-unrelated toxicity of α-GalCer; and (4) in addition to NKT cells, other innate effector cells will be activated by CpG-stimulated PDC, such as NK cells and γδ T cells that may act in concert with NKT cells.
In conclusion, this study identifies a novel immunological activity of CpG ODN: to enhance antigen-specific NKT cell responses. We propose a novel function of PDC that is to translate information about the detection of virus (for which PDC are specialised) to an appropriate T cell effector response not only via type I IFN secretion, but also via a direct cell-to-cell contact irrespective of its ability to present antigen, and distinct from the type of costimulatory activity of MDC.
Materials and methods
ODN, cytokines and blocking antibodies
Completely and partially phosphorothioate-modified ODN were provided by Coley Pharmaceutical Group, Wellesley, MA (small letters: phosphorothioate linkage; capital letters: phosphodiester linkage 3′ of the base; bold: CpG-dinucleotides): ODN 2006: 5′-tcgtcgttttgtcgttttgtcgtt-3′ 53; ODN 2216: 5′-ggGGGACGATCGTCgggggG-3′ 52; ODN 2243 (GC-control to ODN 2216): 5′-ggGGGAGCATGCTCgggggG-3′. ODN were negative for endotoxin in the LAL assay (LAL assay BioWhittaker, Walkersville, MD; lower detection limit 0.1 EU/mL). CpG ODN were used at a concentration of 6 μg/mL or 3 μg/mL as indicated. Recombinant IFN-α (2a) was from PBL (New Brunswick, NJ). For blocking of IFN-α and IFN-β function, a combination of polyclonal rabbit anti-human IFN-α (5000 neutralising U/mL) and rabbit anti-human IFN-β (2000 neutralising U/mL) antibodies together with 20 μg/mL of a mouse anti-human IFN-α/-β receptor chain 2 antibody were used (PBL). IL-12 was neutralised by anti-human IL-12 mAb (10 μg/mL; R&D Systems). α−GalCer was used at 100 ng/mL. In some experiments, anti-CD3-loaded microbeads (M-450 CD3, Dynal Biotec, Norway) were used as model antigen at a concentration of 2.5 × 105 beads/mL.
Preparation, isolation and culture of cells
Human PBMC were isolated from buffy coats provided by the blood bank of the University of Greifswald, Germany as described 65. Blood donors were 18- to 65-year-old healthy men and women who were tested to be negative for HIV, hepatitis B virus, and hepatitis C virus. NKT cells were enriched by positive selection using magnetically activated cell sorting as follows: PBMC were incubated with FITC-labelled anti-Vα24 antibodies (20 μL/4 × 108–8 × 108 cells in 300 μl volume; Immunotec/Coulter, France) followed by labelling with anti-FITC microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) and separation using two consecutive separation columns (LS then MS columns; Miltenyi Biotec) (purity of Vα24+ T cells: 70–96%). Untouched CD8 T cells were isolated from PBMC by depleting other cell types (CD8 T cell isolation kit, Miltenyi Biotec). PDC and MDC were isolated using the BDCA-4 and BDCA-1 DC isolation kits, respectively, according to the manufacturer's protocol (Miltenyi Biotec). Purity of isolated PDC was between 63% and 98%. Purity of isolated MDC was higher than 95%. For some experiments PDC- or MDC-depleted PBMC were used. PDC-depleted PBMC contained less than 0.1% PDC. MDC-depleted PBMC contained less than 0.03% MDC. Viability of cells was determined by trypan blue exclusion. Isolated NKT cells (2 × 104–5 × 104 Vα24+ cells/200 μL) were co-incubated with autologous PBMC (2 × 106 cells/mL) in round-bottom 96-well plates in the presence or absence of 6 μg/mL CpG ODN in RPMI culture medium (Biochrom) supplemented with 10% heat-inactivated (56°C, 1 h) FCS (Biochrom), 1.5 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin (all from Sigma, Munich, Germany). Where indicated, α-GalCer was added at 100 ng/mL. ODN 2216-conditioned medium (ODN 2216-CM) was prepared by collecting the cell-free supernatant of purified PDC (2 × 104–5 × 104 cells/200 μL) after stimulation with CpG ODN (3 μg/mL) for 48 h, or, where indicated, the cell-free supernatant of PBMC (3 × 106 cells/mL) after stimulation with CpG ODN 2216 (3 μg/mL) for 24 h. For some experiments, a transwell culture system allowing only exchange of soluble factors but no cell-to-cell contact (0.2 μm, Anapore membrane device; Nunc, Roskilde, Denmark) was used. Enriched NKT cells (2 × 104–5 × 104 Vα24+ cells in 4 × 105 PDC-depleted PBMC in 500 μL medium) in the lower chamber were co-cultured with purified PDC (3 × 104 cells in 500 μL medium) in the upper chamber. For OX40L inhibition experiments, PDC were preincubated with anti-human OX40L/TNFSF4 mAb (clone: 159403; R&D Systems) or an IgG1 control antibody (Sigma) for 30 min prior to addition of T cells.
At the indicated time points, cells were harvested and surface antigens were analysed by flow cytometry as previously described 66. Anti-CD3PerCP, anti-CD69PE, anti-CD1dPE, lin cocktailFITC, anti-CD123PE, anti-CD11cAPC, anti-CD80PE, anti-CD83FITC and anti-HLA-DRPerCP were purchased from BD Biosciences (Heidelberg, Germany). Anti-Vα24FITC and anti-Vβ11-biotin were from Immunotech/Coulter (France). For Vβ11 staining, cells were incubated with Vβ11-biotin followed by staining with streptavidin-allophycocyanin (BD Biosciences). Flow cytometric data were acquired on a Becton Dickinson FACSCalibur equipped with two lasers (BD Biosciences). Analysis was performed on viable cells. Data were analysed using CellQuest software (Becton Dickinson, Heidelberg, Germany).
Detection of cytokines by ELISA
IFN-γ and IL-4 were measured in the cell-free supernatants using commercially available ELISA kits according to the manufacturer's protocol (BD OptEIA ELISA sets from BD Pharmingen; range IFN-γ ELISA: 5–300 pg/mL; range IL-4 ELISA: 8–500 pg/mL).
Real time RT-PCR
Purified cell populations were cultured for 3 h in RPMI 1640 with 10% FCS in the absence of stimulation. As detailed earlier 43, total RNA was extracted, reverse transcribed using AMV-reverse transcriptase (RT) and oligo-(dT) primer and real time PCR was performed using TLR-specific primers. High-Pure-RNA-Isolation Kit and First Strand-cDNA-Synthesis Kit for RT-PCR were purchased from Roche (Mannheim, Germany). The data of two independent analyses for each sample and parameter were averaged. The copy number of the different TLR is presented as number of transcripts per 103 copies of the housekeeping gene cyclophilin-B.
Data are expressed as means ± SEM. Statistical significance was determined by the paired Student's t-test. Differences were considered statistically significant for p<0.05. Statistical analyses were performed using Stat-View 4.51 software (Abacus Concepts, Calabasas, CA).
This study was supported by grants SFB 571, DFG HA 2780/4–1, and the Dr. Mildred Scheel Stiftung 10–2074 to G. H. and a grant of the MMW to S. R. This work is part of the thesis of D. P. at the University of Munich.
- 6Natural T cells in the human liver: cytotoxic lymphocytes with dual T cell and natural killer cell phenotype and function are phenotypically heterogenous and include Valpha24-JalphaQ and gammadelta T cell receptor bearing cells. Hum. Immunol. 1999. 60: 20–31.
- 12[Development of KRN7000, derived from agelasphin produced by Okinawan sponge]. Nippon Yakurigaku Zasshi 1997. 110 Suppl 1: 63P–68P.