SEARCH

SEARCH BY CITATION

Keywords:

  • CD4+ T cell;
  • dendritic cell;
  • dexamethasone;
  • protease-activated receptor-2;
  • tissue factor

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References

The precise function of tissue factor (TF) expressed by dendritic cells (DC) is uncertain. As well as initiating thrombin generation it can signal through protease-activated receptor 2 (PAR-2) when complexed with factor VIIa. We investigated the expression and function of TF on mouse bone marrow (BM) -derived DC; 20% of BM-derived DC expressed TF, which did not vary after incubation with lipopolysaccharide (LPS) or dexamethasone (DEX). However, the pro-coagulant activity of DEX-treated DC in recalcified plasma was 30-fold less than LPS-treated DC. In antigen-specific and allogeneic T-cell culture experiments, the TF on DEX-treated DC provided a signal through PAR-2, which contributed to the reduced ability of these cells to stimulate CD4+ T-cell proliferation and cytokine production. In vivo, an inhibitory anti-TF antibody and a PAR-2 antagonist enhanced antigen-specific priming in two models where antigen was given without adjuvant, with an effect approximately 50% that seen with LPS, suggesting that a similar mechanism was operational physiologically. These data suggest a novel TF and PAR-2-dependent mechanism on DEX-DC in vitro and unprimed DC in vivo that contributes to the low immunogenicity of these cells. Targeting this pathway has the potential to influence antigen-specific CD4+ T-cell activation.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References

Initiation and propagation of the coagulation cascade in haemostasis and thrombosis is well characterized and results in the formation of a fibrin clot. Tissue factor (TF) is the physiological initiator of thrombin generation and is constitutively expressed on vascular adventitial cells, leading to the traditional concept that TF forms a ‘haemostatic envelope’ to prevent bleeding from damaged vessels.[1]

Tissue factor can also be induced on inflammatory cells and has been implicated in the pathogenesis of a variety of diseases such as atherosclerosis, malignancy and antibody-mediated rejection.[2] Although some of these processes involve thrombosis, some of the effects of TF are fibrin-independent and involve signalling through protease-activated receptors (PAR), which are cleaved by serine proteases to expose a neo-N-terminal activating tethered ligand.[3] Thrombin can cleave and activate PARs 1, 3 and 4, whereas TF bound to factor VIIa (FVIIa) in a binary complex is capable of activating PAR-2 directly. PAR-2 is also cleaved by a number of other proteases including FXa, trypsin, proteinase 3 and mast cell tryptase.

Tissue factor and PARs are expressed by a number of immune cells including monocytes, macrophages and dendritic cells (DC).[4] On antigen-primed DC, thrombin, via PAR-1, has a profound influence on migration through lymph nodes and subsequent spread of systemic inflammation in murine models of endotoxaemia and infection.[5] Additionally, PAR-2 signalling has been shown to enhance uptake of antigens, trafficking and T-cell activation by DC.[6-10]

In this paper, we investigated the function of TF on mouse bone marrow (BM) -derived DC, comparing DC outgrown under standard conditions with those incubated with either dexamethasone (DEX) or lipopolysaccharide (LPS). Our data suggest that, despite similar expression levels of TF by all three, there is a hierarchy of pro-coagulant activity and TF expressed by DEX-treated DC appears to provide signalling through PAR-2 to sustain the low immunogenicity of these cells. Based on additional data generated in vivo, we speculate that a similar mechanism operates physiologically to limit CD4+ T-cell priming to specific antigens encountered in the absence of an adjuvant.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References

Animals

Six- to eight-week-old BALB/c (H-2d) and C57BL/6 (H-2b) mice were purchased from Harlan Laboratories (Bicester, UK). Two strains expressing transgenic T-cell receptors (TCR) were used to assess antigen-specific responses, chosen because they were available within the Immunology Department. The first, Marilyn, has a transgenic TCR specific for male HY antigen restricted by MHC class II and the second, DO11.10, has a TCR that recognizes an ovalbumin peptide (OVA323–339) restricted by H-2Ad. Both were kind gifts from Drs Jian Guo Chai and Andrew George (Imperial College London, UK). All procedures were performed in accordance with the Home Office Animals (Scientific Procedures) Act of 1986.

Cell isolation and T-cell assays

The BM-derived DC were cultured as described elsewhere.[11] Briefly cells were flushed from the femurs and tibiae of mice and passed through a nylon cell strainer. Red blood cells were lysed using ammonium chloride–potassium buffer. After washing, 1 × 106/ml cells were seeded in RPMI-1640 growth medium plus 5% supernatant from a granulocyte–macrophage colony-stimulating factor (GM-CSF) -producing hybridoma cell line. On day 3, non-adherent cells were discarded and fresh medium was added. In some cultures, either 1 μm DEX (Sigma-Aldrich, Poole, UK)[12] or 1 μg/ml LPS (Escherichia coli serotype 0128:B12) (Sigma-Aldrich) was added on days 5 or 6. The DC were harvested on day 7.

T cells were isolated from the spleen and lymph nodes (mesenteric, inguinal and axillary). Organs were passed through a nylon cell strainer and red blood cells were lysed as above. Splenocytes were incubated with an antibody cocktail supplied by Invitrogen (Carlsbad, CA) containing rat anti-mouse Gr, CD16/32, MHCII and CD8 antibodies for 20 min at 4° before washing and incubation with sheep anti-rat magnetic beads for negative selection according to manufacturer's instructions. The resulting CD4+ T cells were 90–95% pure.

To assess T-cell proliferation against alloantigens, 2 × 105 BALB/c T cells were stimulated with 1 × 104 irradiated C57BL/6 DC in 200 μl complete medium unless otherwise stated. To assess antigen-specific proliferation, 2 × 105 female Marilyn CD4+ T cells were stimulated with 1 × 104 male C57BL/6 DC in 200 μl complete medium. In some assays, rabbit polyclonal anti-TF antibody (American Diagnostica, Stamford, CT) or control rabbit immunoglobulin were added at the start. Proliferation was measured by adding [3H]thymidine on day 4 of culture and harvesting 16–18 hr later to determine T-cell proliferation as assessed by incorporated radioactivity.

Flow cytometric analysis

All flow cytometry was performed on a FACSCalibur flow cytometer and analysed using Cellquest (BD BioSciences, Oxford, UK) or Flojo (Treestar, Ashland, OR) software. For cell surface analysis, the following antibodies were used; rat anti-mouse CD4, CD8, (e-Bioscience, San Diego, CA) FITC-CD80 (Serotec, Kidlington, UK), FITC-CD86 (Becton Dickinson, Oxford, UK); hamster anti-mouse FITC-CD3, FITC-CD11c, FITC-MHC II (e-Bioscience); rabbit polyclonal anti-TF, anti-TFPI (both American Diagnostica), PAR-3, PAR-4 (Santa Cruz Biotechnology, Dallas, TX); mouse anti-PAR-1 (Becton Dickinson), PAR-2 (Santa Cruz Biotechnology). Where appropriate, the following second layers were used: swine anti-rabbit FITC-immunoglobulin (Dako, Glostrup, Denmark); goat anti-rabbit FITC-immunoglobulin, anti-rabbit phycoerythrin-immunoglobulin (Sigma-Aldrich), anti-mouse FITC-IgG (Dako); mouse anti-rat FITC-immunoglobulin (e-Bioscience).Then, 2 × 105 cells were analysed immediately or fixed in 2% paraformaldehyde in PBS and analysed within 3 days.

Intracellular cytokine staining was performed as previously described.[13] Briefly, cells were stimulated with 50 ng/ml PMA (Sigma-Aldrich) plus 500 ng/ml ionomycin (EMD Biosciences, Darmstadt, Germany) for 4 hr, with 10 μg/ml brefeldin A (Sigma-Aldrich) for the final 2 hr. All washes and incubations were carried out in buffer containing 0·5% Saponin (Sigma-Aldrich). Cells were stained with rat anti-interferon-γ (IFN-γ), interleukin-4 (IL-4) or IL-10 (all from BD Pharmingen, Franklin Lakes, NJ, USA)

RNA extraction and RT-PCR

Between 5 × 106 and 1 × 107 cells were washed thoroughly with PBS before RNA was extracted using phenol and chloroform and re-suspended in RNAse-free water (Sigma-Aldrich). RNA was assessed using agarose gel analysis and Quanti-iT Ribogreen RNA reagent and kit (Invitrogen, Paisley, UK). RT-PCR was peformed using reagents from Applied Biosystems (Carlsbad, CA), including primers for PARs 1–4 and β-actin. All PCR products were run on 1% agarose gel.

Clotting assay

Mouse acetone brain extract (Sigma-Aldrich), used as a standardized source of TF and all other reagents were suspended in 50 mm Tris–HCl, 150 mm NaCl and 1 mg/ml human albumin pH 7·4. For test samples, cells were suspended at a concentration of 1 × 107/ml. Serial dilutions of brain extract (in 80 μl) or 1 × 107 cells/ml (80 μl) were mixed in a glass tube with 80 μl phospholipid and 80 μl pooled normal mouse plasma at 37° for 1 min. To start the clotting assay 80 μl 65 mm CaCl2 was added, and, while being continuously agitated, the time for a clot to form in the tube was measured. In some assays, rabbit polyclonal anti-TF antibody or control rabbit immunoglobulin were added at the start. All samples were performed in triplicate. A standard curve generated from the TF in mouse brain extract was used to measure relative TF function in the test samples.

In vivo antigen-specific T-cell sensitization

Cells from lymph nodes and spleen of naive DO11.10 mice were labelled with 2·5 μm carboxyfluorescein succinimidyl ester (CFSE; Cambridge Biosciences, Cambridge, UK) using standard protocols. 1 × 107 cells were injected into the tail vein of BALB/c mice. After 24 hr, the mice were given an intraperitoneal injection of either saline, 250 μg anti-TF antibody,[14] control rabbit immunoglobulin, 25 μg LPS (E. coli serotype 0127:B8) (Sigma-Aldrich) or LPS alone followed by 5 mg whole OVA. After 72 hr, lymph node (cervical, axillary, brachial and inguinal) cells were harvested for analysis of proliferation and IFN-γ, IL-4 and IL-10 production by flow cytometry.

Contact sensitivity responses

On day 0, C57BL/6 mice were sensitized by application of 5% oxazalone in ethanol and acetone (4 : 1, 50 µl) to the shaved abdomen. At the same time mice were injected intraperitoneally with either 100 μl saline control, 1 µm/kg PAR-2 agonist or 100 µm/kg PAR-2 antagonist. Mice were re-challenged on day 5 by applying 1% oxazolone in olive oil and acetone (4 : 1, 10 µl) to the right ear, whereas the left ear was painted with vehicle alone. Ear thickness was measured using a digitial micrometer. Measurements were made on day 5 as a baseline and then 24 and 48 hr following ear painting.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References

TF expression and pro-coagulant function on in vitro cultured DC

C57BL/6 BM-derived DC (> 85% CD11c+) were examined by flow cytometry. Figure 1 illustrates the expression of MHC class II, CD80 and CD86 by cells grown under three different conditions. As expected, DEX-treated DC expressed low levels of MHC II and co-stimulatory molecules whereas LPS-treated DC had higher expression. The DC cultured without either DEX or LPS (‘untreated’) had an intermediate phenotype. Subpopulations of DC from all three conditions expressed detectable TF antigen on their cell surface, representing 15–20% of the CD11c+ cells, with similar population MFI (Fig. 1b–d), suggesting similar levels of TF expression. Similar results were obtained with DC grown from BALB/c mice (data not shown).

image

Figure 1. Flow cytometric analysis of murine bone marrow (BM) -derived dendritic cells (DC) to assess tissue factor (TF) expression on DC at different states of functional maturity. (a) C57BL/6 BM cells were grown in granulocyte–macrophage colony-stimulating factor and DC harvested on day 7 of culture before analysis by flow cytometry with the antibodies as indicated. Cells that were otherwise untreated are compared with those in which dexamethasone (DEX, 1 μm) or lipopolysaccharide (LPS, 1 μg/ml) were added into the culture for the final 48 or 24 hr, respectively; representative of four experiments. DEX-treated (b), untreated (c) and LPS-treated (d) DC were harvested on day 7 of culture and analysed for co-expression of CD11c and TF. Profiles show isotype control and specific binding. Numbers in quadrants represent percentage of positively stained cells within each quadrant; representative profiles from one of four experiments.

Download figure to PowerPoint

In clotting assays using re-calcified mouse plasma, addition of all three DC populations promoted shorter clotting times compared with the ‘no cells’ control, with a hierarchy of clotting times (LPS < untreated < DEX) (Fig. 2a), each of which was prolonged by an inhibitory anti-TF antibody (Fig. 2b) suggesting that there were significantly different levels of pro-coagulant TF present on each type of DC. This was confirmed when the relative TF pro-coagulant activity on each was calculated (Fig. 2c). In these experiments, the time to clot spontaneously in glass tubes (‘no cells’ in Fig. 2a) represents the clotting induced by the (TF-independent) contact-activated pathway. These data indicated that a greater proportion of TF on DEX-treated DC was in the non-pro-coagulant or ‘encrypted’ form. There were no detectable differences in annexin V binding, protein disulphide isomerase or murine tissue factor pathway inhibitor expression by any of the DC used in these assays (data not shown).

image

Figure 2. Analysis of tissue factor (TF) function on murine dendritic cells (DC). (a, b) Clotting assays with re-calcified mouse plasma in presence of dexamethasone-treated (DEX), untreated (UNTX) or lipopolysaccharide-treated (LPS) DC or B16F10 tumour cells known to express high levels of surface TF [positive control in (a)]. The ordinate shows the time to clot in seconds (± SEM). Error bars in some columns are too small to see. In (b) respective DC were incubated with increasing doses of anti-TF (closed triangles) or rabbit control (open triangles) immunoglobulin. (c) Relative TF activity on DEX-treated and LPS-treated DC compared with untreated (UNTX) DC (= 1) calculated from the clotting times from standard curves generated by titrating fixed concentrations of TF from mouse brain extract into the clotting assay. *< 0·01 compared with untreated. (d) C57BL/6 male DEX DCs were incubated with either medium, polyclonal rabbit anti-TF antibody or control rabbit immunoglobulin as indicated and co-cultured with purified Marilyn female CD4+ T cells. The T-cell response was measured by 3H incorporation on day three of culture. Graph shows mean ± SEM of triplicate well. Compared with control immunoglobulin *P = 0·05. (e) Experiment similar to that in (d), except DEX-treated BL/6 DC incubated with BALB/c CD4+ T cells. Compared with control immunoglobulin, *P < 0·002. (f) As (e), except untreated DC used as stimulators. = NS. (g) As (e), except CD8+ T cells used as responders. = NS. All data representative of three experiments.

Download figure to PowerPoint

Biological activity of TF on DEX-treated DC

To investigate whether TF on DEX-treated DC had a biological function, male DCs were co-cultured with naive CD4+ T cells from female Marilyn mice, which possess a transgenic TCR that recognizes the male HY antigen restricted by MHC class II. Compared with DEX-treated DC incubated with isotype control immunoglobulin, DC incubated with inhibitory anti-TF antibody stimulated significantly greater proliferation (Fig. 2d). These experiments were repeated with allogeneic DC and wild-type CD4+ T cells, with similar results (Fig. 2e), suggesting that the effect was not due to antigen processing. The same antibody did not enhance the near-maximal proliferation induced by untreated DC (Fig. 2f) and had no impact on proliferation of CD8+ T cells induced by DEX-treated DC (Fig. 2g). These results indicated that inhibiting TF on DEX-DC specifically enhanced proliferation by CD4+ T cells and suggested that the TF on DEX-DC was acting to limit the capacity of these cells to induce CD4+ T-cell proliferation.

The TF–FVIIa complex can cleave and activate PAR-2 signalling.[15] Analysis of PAR expression by DEX-DC revealed transcripts for all PARs (Fig. 3a), but only PAR-2 and PAR-4 were detectable on the cell surface by flow cytometry (Fig. 3b), which may reflect known differences in the importance of mRNA stabilization for expression of different PARs.[16] To address whether the effect of TF was due to signalling through PAR, agonist or antagonist peptides were added to DEX-DCs before co-culture with CD4+ T cells. Figure 3(c,d) shows that, in contrast to when PAR-1 or PAR-4 agonists were used, a selective PAR-2 agonist inhibited the effect of the anti-TF antibody, indicating that PAR-2 activation was able to compensate for the inhibition of TF when antibody was present. In contrast, a PAR-2 antagonist enhanced the proliferative capacity of DEX-treated DC to the same extent as the anti-TF antibody (Fig. 3e), suggesting that it had the same functional effect as the antibody. These data are consistent with the hypothesis that the (mostly) encrypted TF on DEX-treated DC influences the capacity to stimulate T cells through activation of PAR-2.

image

Figure 3. Experiments to determine protease-activated receptor (PAR) expression on mouse dendritic cells (DC) and interaction of tissue factor (TF) with PAR-2 (a) RT-PCR analysis of RNA extracted from dexamethasone (DEX) -treated (D), untreated (U) or lipopolysaccharide (LPS) -treated (L) DC for expression of PAR-1 to PAR-4. Control lane (N) contains RNA from DEX-treated cells without reverse transcriptase. β-actin expression used as positive control for RNA integrity. (b) Flow cytometric analysis of DEX-treated DC for expression of PAR-1 to PAR-4 (open profiles). Shaded profiles show istotype control binding. (c, d) DEX-treated DC were incubated with medium alone (medium), anti-TF antibody, or control rabbit immunoglobulin as indicated with or without additional agonist peptides to activate PAR-1 and PAR-4 (c) or PAR-2 (d), before co-culture with purified BALB/c CD4+ T cells. Graphs show mean [3H]thymidine incorporation (± SEM) of triplicate wells on day 5 of culture. Representative of three experiments. = NS all comparisons. (e) As (d) but with additional control of DEX-treated DC incubated with a selective PAR-2 antagonist. Representative of three additional experiments. *P < 0·05

Download figure to PowerPoint

Inhibition of TF/PAR-2 signalling enhances T-cell priming in vivo

To assess whether any of this was relevant in vivo, lymph node and spleen cells from naive DO11.10 mice were isolated and labelled with CFSE. T cells from these mice possess a transgenic TCR that recognizes OVA323–339 restricted by H-2Ad.17 The labelled cells were injected intravenously into naive BALB/c mice 24 hr before intraperitoneal sensitization with whole OVA. After 72 hr lymph nodes from the mice were harvested and examined by flow cytometry. As shown in Fig. 4, compared with mice receiving control IgG, there were increased numbers of DO11.10 cells producing IFN-γ in mice given an anti-TF antibody, the frequency of which approached that seen in mice when LPS was used as adjuvant. Proliferation of these IFN-γ-producing CD4+ cells was also increased compared with when OVA was administered with the control, but these data just failed to reach statistical significance. No differences in IL-4 production were noticed. No TF was found on CD4+ or CD8+ T cells isolated from spleen and lymph nodes of these mice as described (data not shown), indicating that the anti-TF antibody was unlikely to be acting directly on the T cells and consistent with it having an influence on T-cell priming via DC.

image

Figure 4. In vivo experiments demonstrating potential physiological role for tissue factor (TF)/protease-activated receptor 2 (PAR-2) during sensitization. (a) 1 × 107 carboxyfluorescein succinimidyl ester (CFSE) -stained lymph node and spleen cells from naive DO11.10 mice were injected intravenously into naive BALB/c mice. 24 hr later mice were immunized with 5 mg whole ovalbumin (OVA) given with either an anti-TF antibody (250 μg) or control immunoglobulin or 25 μl lipopolysaccharide (LPS; positive control). Negative control mice received no OVA or LPS only. After 72 hr the lymph node cells were harvested from individual mice and analysed for proliferation and cytokine production [interferon-γ (IFN-γ) and interleukin-4 (IL-4)]. Representative profiles from one to nine mice from each group are shown. (b) Experiment performed exactly as in (a). The number of IFN-γ-producing CFSE+ CD4+ T cells were enumerated from each mouse given OVA plus polyclonal immunoglobulin control (open bar), anti-TF (striped bar) or LPS (filled bar). Fold increase was calculated as the ratio of this number for each mouse to the equivalent number of CD4+ T cells enumerated for mice given LPS alone. Mean and standard error were calculated for each group and are plotted. (c, d) Contact hypersensitivity was induced by sensitization with 5% oxazolone on the shaved abdomen of C57BL/6 mice. After 5 days the mice were re-challenged with vehicle alone on the left ear and oxazolone on the right. Graphs show difference in ear thickness between right and left ears 24 hr (c) and 48 hr (d) after re-challenge (= 12). Before initial sensitization, mice received either a PAR-2 antagonist, PAR-2 agonist or normal saline control IP. Data are expressed as an increase in ear thickness. Statistical analyses performed using two-way analysis of variance with Bonferroni's post-test (*P < 0·05 at both time-points compared with saline).

Download figure to PowerPoint

To further investigate, a model of DC-dependent delayed-type hypersensitivity was used. For these experiments, a specific PAR-2 antagonist was used instead of the anti-TF antibody, so that only the signalling via PAR-2 was inhibited. C57BL/6 mice were injected with a PAR-2 antagonist peptide immediately before oxazolone sensitization on the abdominal skin. Five days later mice were re-challenged with oxazolone and antigen-specific swelling was calculated. As shown in Fig. 4(c,d), mice given a PAR-2 antagonist at the time of sensitization developed significantly greater antigen-specific swelling compared with control mice, consistent with the hypothesis that constitutive PAR-2 signalling modulates T-cell priming.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References

Expression of TF has been previously defined on human DC in lymphoid follicles,[17] DC derived from monocytes[4, 18] and on murine DC.[5] In this report, we confirm TF expression by 10–15% of mouse BM-derived DC and show that this proportion and the level of TF expression were not influenced by addition of DEX or LPS in the final 2 days of culture. However, the pro-coagulant activity was significantly different on these cell populations, with DEX-treated DC having approximately fivefold less functional TF activity compared with untreated DC and 30-fold less than that of LPS-treated DC (Fig. 2). These data suggest that DC have a mechanism to regulate the functional activity of TF, in line with their state of maturity, such that TF on immature DC exists predominantly in an encrypted or non-pro-coagulant state.

The molecular basis of TF encryption/decryption remains enigmatic. Potential mechanisms for decryption include an increase in the amount of pro-coagulant phosphatidylserine expressed on the cell membrane, and the conversion of Cys186-Cys209 disulphide bonds by protein disulphide isomerase.[19] We assessed the levels of both of these on the surface of our DC and found no differences (data not shown). Tissue factor pathway inhibitor can also regulate the differential functions of TF,[20] but there were no differences in the levels of murine tissue factor pathway inhibitor on all three types of our DC (data not shown), so the mechanism by which TF was regulated on these DC remains unclear

Encrypted TF/FVIIa is known to signal through PAR-2.[15, 21] Using an inhibitory anti-TF antibody, and a combination of PAR-2 agonists and antagonists, our data indicate that the encrypted TF on DEX-treated DC provides a signal through PAR-2 that suppresses CD4+ T-cell-proliferative responses, so that the anti-TF antibody or PAR-2 antagonist significantly enhanced CD4+ T-cell proliferation after both allogeneic and antigen-specific stimulation. In the presence of the anti-TF antibody these responses could be re-suppressed by pre-incubation of DC with the PAR-2 agonist. This mechanism was not apparent when the anti-TF antibody was used with untreated or LPS-treated DC, suggesting either that it was not operational on these cells, or that the effect was masked by the near maximal proliferative responses induced by these cells.

In vivo, the anti-TF antibody enhanced the priming of naive antigen-specific transgenic T cells to whole OVA (administered without an adjuvant), as shown by increased production of IFN-γ and a trend towards increased proliferation. The magnitude of the effect of anti-TF antibody was approximately 50% of control LPS adjuvant, indicating the potency of the TF-mediated suppression of priming in this system. Additionally, the PAR-2 antagonist given to mice at the time of dermal priming to oxazolone enhanced the subsequent recall response, evident by a significant increase in the ear swelling on secondary challenge with the antigen. In these experiments, a PAR-2 agonist given during the priming phase had no significant impact on recall responses. These data suggest that the mechanism we have defined in vitro has a functional relevance in vivo.

The effects of PAR-2 signalling on the DEX-treated DC we report here are different to those previously reported by others. Fields et al., described that PAR-2 signals were required for DC maturation during GM-CSF-induced outgrowth from BM cells,[6] although these data have since been challenged.[8] A number of studies have concluded that PAR-2 signalling enhances antigen uptake and trafficking by murine DC,[8] and increases T-cell activation in response to specific antigens.[9, 10, 22] In trying to reconcile our data with these, one obvious difference is that we have focused on DEX-treated immature DC; the mechanism we describe appears irrelevant in mature DC. It may also be relevant that others have compared responses in wild-type and PAR-2-deficient mice and have addressed the effect of proteases other than TF/FVIIa.

In summary, we report that the activity of TF, expressed on a subset of murine DC, varies with maturation state, such that TF on LPS-treated DC is pro-coagulant whereas on DEX-treated DC it appears to provide a tonic signal through PAR-2 that maintains an aspect of the low-immunogenic phenotype that is characteristic of these cells. This mechanism appears relevant in vivo, partially suppressing priming and subsequent recall responses. This novel role for TF/PAR signalling augments our understanding of DC function and highlights that manipulating this pathway might influence antigen-specific responses.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References

SS received an NKRF (KRUK) Clinical Fellowship to perform this work. The work was additionally funded by The Medical Research Council, UK (award refs. G0401591 and G0801965) and the Garfield Weston Foundation. AD also receives financial support from the British Heart Foundation, Guy's and St Thomas’ Charity, and the GSTT Kidney Patients Association. We acknowledge support from the National Institute for Health Research (NIHR) Biomedical Research Centre based at Guy's and St Thomas’ NHS Foundation Trust and King's College London. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosures
  9. References