Autologous mixed leukocyte reaction
Current literature suggests that T cells recognizing antigen on mature dendritic cells (DC) differentiate into effector T cells whereas tolerance is induced when antigen is presented by immature DC. We investigated the consequences of the interactions between immature or lipopolysaccharide-matured DC and CD4pos T lymphocytes in absence of foreign antigen. While immature DC did not induce significant CD4pos T cell activation, we observed that a significant fraction of CD4pos T cells cultured with mature autologous DC displayed phenotypic features of activation and produced IL-2, IFN-γ, IL-10 and TGF-β. Furthermore, CD4pos T lymphocytes primed by mature, but not immature, autologous DC acquired regulatory properties. Indeed, when added to an allogeneic mixed leukocyte reaction, they suppressed the response of alloreactive T lymphocytes to the priming DC while responses to third-party stimulators were spared. The generation of CD4pos T cells with regulatory function by autologous stimulation did not require the presence of natural CD4posCD25pos regulatory T cells. In addition, the acquisition ofregulatory function by CD4posCD25neg T cells stimulated by autologous mature DC was accompanied by the induction of FOXP3 expression. Our data suggest that during inflammatory conditions, presentation of self antigens by mature DC to autologous T lymphocytes could contribute to the generation of regulatory mechanisms.
DC are professional APC which are involved both in the induction and the regulation of immune responses 1. In the presence of microbial products interacting with "pattern recognition receptors" 2 and/or in the presence of "danger signals" present during inflammatory process 3, DC mature and acquire the costimulatory signals necessary to induce the activation of T cells specific for the presented antigen. On the other hand, in steady state conditions, immature DC would be specialized for tolerance induction 1. Although this concept has been validated both in vitro4, 5 and in vivo6–8, it is not clear whether it also applies to the presentation of self-antigens to a normal T cell repertoire. We were therefore interested to characterize the phenotype and function of T lymphocytes cocultured with immature or mature autologous DC in the absence of exogenous antigen. Indeed, there is evidence that, in this setting, DC present peptides derived from self proteins 9, mostly derived from apoptoticcells 10, 11. This in vitro model is therefore relevant to mimic the fate of T cells responding to tissue-derived self antigens released during an infection or a trauma.
The response of T cells to autologous stimulators was previously addressed in studies on the so-called autologous mixed leukocyte reaction (AMLR). T cells activated in AMLR were shown to display two classical attributes of an immune phenomenon: memory and specificity 12. Different studies also showed that AMLR led to the generation of T lymphocytes with suppressive properties 13, 14. Although most previous studies used as stimulators unpurified APC, some reports demonstrated that DC were unique in their ability to trigger AMLR 15, 16. Herein, we analyzed the influence of DC maturation on the fate of autologous T cells focusing on their phenotype, cytokine profile and regulatory function.
2.1 LPS-matured autologous DC induce T cell blast transformation and proliferation
To avoid exogenous antigen presentation to autologous T cells, all the experiments were conducted in medium containing autologous plasma, including DC generation. Fig. 1 illustrates the phenotype of DC generated in such conditions. In the absence of maturating stimuli, DC displayed substantial levels of HLA-DR, CD40 and B7–2 molecules and very low levels of CD80, CD83 and CD14. In order to induce DC maturation, we used a common component of Gram-negative bacterial wall, LPS 17, 18. As expected, LPS induced the up-regulation of the expression of MHC class II, costimulatory and CD83 molecules (Fig. 1).
When T cells were cocultured with autologous DC, we observed that T cell blast formation (Fig. 2A) and proliferation (Fig. 2B) occurred in the presence of both immature and mature DC but to a much higher extent in the presence of mature DC. Blastic T cell transformation assessed by FACS analysis of size and granularity parameters was 5.7±1.0% for T cells cultured with mature DC, 1.6±0.2% when cultured with immature DC and 0.7±0.1% when cultured alone [mean ± SEM of seven independent experiments on different donors, p≤0.005 when blastic transformation of T + DC (LPS) was compared to T cells alone or T + DC].
In order to exclude that CD4pos T cell activation by mature autologous DC might represent the reactivation of memory CD4posCD45ROpos T cells in response to minute amount of foreign antigen-derived peptides presented by DC, we analyzed the proliferation of total CD4pos T cells or after depletion of either CD45RApos or CD45ROpos T cells. As illustrated in Fig. 2C, naive CD4posCD45RApos T lymphocytes were the major responding population to mature autologous DC.
2.2 CD4pos T lymphocytes activated by mature autologous DC acquire the phenotype of effector cells
We next looked at T cell surface molecules expression that could give indication on the function and homing characteristics of CD4pos T cells activated by mature autologous DC. As illustrated in Fig. 3, when compared to unstimulated T cells (Fig. 3A), T cells activated by mature autologous DC (Fig. 3B) up-regulated their expression of CD45RO, HLA-DR, CD69, CD25 whereas they down-regulated CCR7 expression. Very few acquired CTLA-4, GITR and CD103 expression. When gating the analysis on T cell blasts (Fig. 3C), we observed that the phenotypic changes mainly affected this population.
2.3 Cytokine profile of CD4pos T lymphocytes activated by mature autologous DC
When looking at cytokine production during the primary stimulation (Fig. 4), we could detect IFN-γ both at mRNA (Fig. 4C) and protein level (Fig. 4A) when T cells were cocultured with mature DC. IFN-γ protein or mRNA could not be detected when T cells were cocultured with immature DC or left alone. IL-2 mRNA was detected at significant levels only when T cells were cultured with mature DC (Fig. 4B). IL-2 could not be detected by ELISA, probably reflecting IL-2 consumption. Kinetic experiments demonstrated that IL-2 and IFN-γ mRNA content reached a peak after 96 h during primary stimulation of T cells by mature autologous DC. Neither IL-10 nor IL-5 could be detected in the supernatant or by real-time RT-PCR in any of the conditions tested.
To better characterize T cells activated by mature autologous DC, such T cells were restimulated by mature DC. We observed that a priming occurred during primary stimulation with mature DC as preactivated T cells proliferated during the restimulation with a kinetic characteristic of a secondary reaction (Fig. 5A). The cytokine profile of T cells primed by mature DC and restimulated indicated that those cells produced significant levels of IL-10 in addition to IFN-γ (Fig. 5B). mRNA content analysis confirmed an increase of mRNA for IL-10 and IFN-γ (Fig. 5C and F) and revealed in addition the synthesis of IL-2 and TGF-β mRNA (Fig. 5D and E). The mRNA content during the secondary stimulation of T cells by mature autologous DC peaked at 6 h for IL-2 and 24 h for IFN-γ, IL-10 and TGF-β.
2.4 CD4pos T cell activation by mature autologous DC requires MHC class II and costimulatory molecules
Recent observations have suggested that T cell signaling can occur in the absence of MHC class II-TCR interaction leading to increased T cell survival 19. Other studies, however, have established that such interaction is necessary for T cell activation in autologous MLR 16, 20. To determine whether T cell activation by mature autologous DC also required MHC class II – TCR interaction, mAb against MHC class II molecules was added to the coculture of mature DC and T cells. We also tested the requirement of costimulatory molecules for such activation. Results are depicted in Fig. 2D and show that the addition of either anti-MHC class II or anti-CD80 and CD86 mAb to the coculture of autologous mature DC and T cells, lowered T cell proliferation to the level of unstimulated T cells.
2.5 CD4pos T lymphocytes primed by mature autologous DC exert suppressive function
Previous studies have suggested that self-MHC class II-reactive T cells can exert suppressive functions 13, 14. We observed that control fresh CD4pos T lymphocytes and CD4pos T lymphocytes primed by immature autologous DC did not exert any significant effect on the proliferation of allogeneic T lymphocytes activated by DC generated from the same donor as for AMLR (Fig. 6A). In contrast, T cells primed by autologous mature DC inhibited such T cell activation in a dose-dependent fashion (Fig. 6A). At a ratio of one added T cell for ten responder T cells, the suppression was already 40%. At a ratio of one added T cell for two responder T cells, suppression was maximal and very reproducible with a mean inhibition of 77% in seven independent experiments performed on different donors (Fig. 6B). When looking at cytokine production, we observed that addition of T cells primed by mature DC resulted in a profound inhibition of IFN-γ and IL-2 production (Fig. 6C) while IL-5 and IL-10 were not significantly affected (data not shown). To investigate whether the suppressive function of T cells primed by mature autologous DC required their reactivation, we compared their suppressive function in an allogeneic MLR conducted either with DC from the same donor as the one used for the autologous reaction or with DC from an unrelated donor. As shown in Fig. 6D, T cells primed by mature autologous DC needed to be reactivated by DC from the same donor to exert suppression, conferring antigen specificity to the regulatory mechanism induced.
2.6 Suppressive function of CD4pos T cells primed by mature autologous DC is affected neither by the addition of growth factors nor by neutralization of suppressive cytokines
As a first approach to study the mechanisms involved in the suppressive action of CD4pos T cells primed by mature autologous T cells, we tested whether supernatants of T cells primed by mature autologous DC could also exert some regulatory effect. We observed that the addition of such supernatants to the culture medium at a ratio of 1:1 (v/v) induced a suppressive effect of 33, 46 and 50% as measured on day 5 of the read-out allogeneic MLR in three independent experiments. In addition, we observed that supernatants did not modify the kinetics of the allo-MLR (Fig. 7). We then repeated the read-out allogeneic MLR in presence of rIL-2, rIL-15 and rIL-7 added alone or together at concentrations known to be supraoptimal for the induction of T cell growth. As shown in Fig. 8A, the addition of growth factors did not restore the proliferative response, ruling out growth factor consumption as a dominant mechanism. We also tested the effect of the addition of z-vad, a caspase inhibitor (Fig. 8A), of neutralizing antibodies to IL-10 and TGF-β, and of antibodies to CTLA-4 (Fig. 8B). As none of these agents prevented the suppressive activity, it appears that neither activation-induced cell death nor other classical regulatory mechanisms are involved in this in vitro system.
2.7 CD4posCD25neg T cells exert suppressive function after stimulation by mature autologous DC and acquire FOXP3 expression
Since CD4pos CD25pos T cells are naturally occurring regulatory T cells with high affinity for self antigens 21, we considered a possible role for these cells in our model. As illustrated in Fig. 9A, depletion of CD4posCD25pos T cells before activation by autologous DC did not affect the suppressive function of CD4pos T cells, demonstrating that CD4posCD25neg T cells can acquire suppressive function independently of the presence of naturally occurring CD4posCD25pos regulatory T cells. FOXP3 is now recognized as a master control gene for the development and function of regulatory T cells 22. We therefore analyzed FOXP3 mRNA expression before and after stimulation of CD4posCD25neg T cells by mature autologous DC, using naturally occurring CD4posCD25pos T cells from the same donors as positive control. As shown in Fig. 9B, a clear induction of FOXP3 mRNA expression was observed after autologous stimulation of CD4posCD25neg T cells.
In this study, we observed that a fraction of CD4pos T cells that encounter mature autologous DC in the absence of exogenous antigen proliferate, secrete cytokines including IL-10 and TGF-β, and acquire regulatory properties. In agreement with previous reports 16, 20, this type of autologous T cell stimulation required TCR-self MHC class II interaction. Although the precise nature of the peptide(s) recognized by T cells in this setting is unknown, it was previously established that MHC class II molecules bind self peptides in absence of exogenous antigens 9 and that expansion of self-reactive MHC class II-restricted T cells is driven by auto-antigens derived from apoptotic cells present in the coculture [10;11]. The fact that regulatory T cells generated during the autologous reaction proliferated and produced cytokines of the effector type is consistent with several recent observations 23–27.
The observation that regulatory T cells were induced only when autologous DC were in a mature state might appear as a paradox since most in vivo and in vitro models using non-self antigens conferred a capacity of tolerogenic cells to immature DC 4–8. This suggests that TCR avidity for the antigen dictates the conditions required for the emergence of regulatory mechanisms. For the low avidity self-reactive T cells that escape thymic selection 28, differentiation into regulatory T cells might require optimal costimulation in face of weak TCR signaling provided by self-antigen recognition, as described in other models of partial T cell activation (reviewed in 29). This hypothetical scheme would fit with the observations of Inaba et al. 30 showing the presence of mature DC in lymph nodes at the steady state and the recent publication of Lohr et al. indicating that mature DC constitutively expressing B7 molecules are required for self tolerance maintenance 31.
Although naturally occurring CD4posCD25pos regulatory T cells are known to be self-reactive 21, they did not play a dominant role in the generation of suppressor cells induced by mature autologous DC in our system. Indeed, their depletion did not prevent the generation of regulatory T cells. In addition, CD4posCD25neg acquired FOXP3 expression following autologous stimulation. This is consistent with the very recent report of Walker et al. 32 who showed that activation of CD4posCD25neg T cells by a combination of anti-CD3 and anti-CD28 mAb leads to the emergence of FOXP3 expressing regulatory T cells. We therefore suggest that interaction between self-reactive CD4pos CD25neg T cells and mature autologous DC might be a physiological pathway to generate FOXP3-expressing regulatory CD4pos T cells in the periphery. The regulatory T cells generated via this peripheral pathway could join the pool of natural CD4posCD25pos regulatory T cells which differentiated in the thymus.
As in other in vitro systems previously described, the molecular mechanisms of suppression remain elusive 33. Although our regulatory T cells produced IL-10 and TGF-β and their supernatants exerted suppressive activity, we could not demonstrate a direct role of these inhibitory cytokines. Likewise, CTLA-4 blockade had no significant effect. Whatever its mechanism, this suppressive pathway is rather robust since it resisted the addition of a cocktail of growth factors including IL-2, IL-15 and IL-7, excluding a simple consumption of growth factors by the suppressor T cells.
Interactions of mature CD4pos T lymphocytes with self-peptide-MHC complexes in the periphery appear to be essential for the homeostasis of the immune system. Indeed, such interactions are critical to ensure naive T cells survival and maintenance of themature TCR repertoire 34. In addition to their contributions to T cell homeostasis, low affinity interactions with self-peptide-MHC class II complexes may play a role in immune regulation. Indeed, such complexes were shown to favor immunological synapse formation and induce partial TCR phosphorylation, resulting in an increased sensitivity of T lymphocytes to foreign antigen 35. In accordance to a model proposed by Taams et al. 36, our data suggest that during inflammatory conditions, presentation of self antigens by mature DC to autologous T cells could contribute to the generation of suppressor mechanisms. This process might be critical to prevent autoimmune pathology when self antigens are released in the context of danger signals promoting dendritic cell activation, such as during infections or ischemia-reperfusion injury. As self-reactive regulatory T cells down-regulate their expression of the lymph node homing receptor CCR7, their reactivation would occur at inflammatory sites, thereby allowing targeted suppression at the site of tissue damage. In the course of infection, this fail-safe mechanism would become dominant in the late phase of the immune response when thepathogen is eradicated. Indeed, in the early invasive phase of infection, microbial antigens are in great excess so that self-activation would be dominated by the immune response to the infectious antigens.
The generation of suppression pathways by autologous mature DC is consistent with several observations made in experimental models in vivo. Indeed, Morel et al. 37 showed that administration of unpulsed syngeneic mature DC to NOD mice significantly protected against diabetes development. Likewise, Bhandoola et al. 38 demonstrated that self-MHC class II expression is required to prevent hyperactivation of self-reactive CD4+ T cells. We can also speculate that the prevention of autoimmune disorders by certain microbial products or proinflammatory agents involves this type of mechanisms 39–41. In humans, the correlation between deficient AMLR and the occurrence of autoimmune disease 42, 43 as well as the epidemiological association between poor hygienic condition and protection against type I diabetes and multiple sclerosis 44 could be explained in a similar manner.
In conclusion, we suggest that T cell activation by mature autologous DC could be involved in the maintenance of self-tolerance and the regulation of natural immune responses.
4 Materials and methods
4.1 Culture medium and reagents
Culture medium consisted of RPMI 1640 (BioWhittaker, Verviers, Belgium) supplemented with 2 mM L-glutamine (BioWhittaker), 50 μM 2-ME (Merck-Eurolab, Leuven, Belgium), 1% nonessential amino acids (BioWhittaker). The 2% autologous plasma was added to culture medium for DC generation and 5% autologous plasma for DC-T cell coculture. RhIL-4 was a kind gift from Kris Thielemans (VUB, Brussels, Belgium) and rhGM-CSF was obtained from Novartis (Basel, Switzerland). LPS from E. coli (serotype 0128: B12) was obtained from Sigma Chemicals (Bornem, Belgium). RhIL-2, rhIL-15 and rhIL-7 were purchased from R&D System (Abingdon, GB). The broad-range caspase inhibitor z-VAD was purchased from Enzyme systems products (Livermore, CA).
For immunostaining, we employed the following FITC-conjugated mAb against: CD14 (Becton Dickinson, Erembodegem, Belgium), CD40 (Biosource International, Camarillo, CA), HLA-DR (Becton Dickinson), CD25 (Becton Dickinson), CD45RA (Becton Dickinson), CD103 (Dako, Denmark); the following PE-conjugated mAb against: CD86 (PharMingen, Erembodeghem, Belgium), CD80 (Becton Dickinson), CD83 (Immunotech, Marseille, France ), CD45RO (Becton Dickinson), CD69 (Becton Dickinson), CTLA-4 (PharMingen), GITR (R&D), CCR5 (PharMingen); and unconjugated mAb against CCR7 (R&D) followed by rabbit anti-mouse IgG (Dako). For neutralization experiments, 10 μg/ml unconjugated mAb against IL-10 (R&D), TGF-β1,-β2,-β3 (R&D), CTLA-4, CD86 and CD80 (Becton Dickinson) and anti-HLA-DR, DP, DQ (Dako) were used.
4.3 Cell preparations
Monocyte-derived DC were generated from PBMC using a previously described protocol with slight modifications 45. Briefly, adherent cells were cultured in 3 ml of culture medium containing GM-CSF (800 U/ml), IL-4 (200 U/ml) and 2% autologous plasma in 6-well plates. After 6 days of culture, 100 ng/ml of LPS was added overnight to induce DC maturation. DC were then harvested, washed and used for subsequent experiments. DC purity was more than 90% DC as assessed by morphology and FACS analysis. CD4pos T lymphocytes were purified from fresh non-adherent cells using MACS columns and a negative selection CD4pos T cell isolation kit (Miltenyi Biotech, Auburn, CA), the whole isolation procedure was conducted at 4°C; they were more than 90% pure as assessedby FACS analysis. CD4posCD45RApos purification was performed by negative selection using CD4posCD45ROpos microbeads (Miltenyi Biotech). Depletion of CD4posCD25pos T cells was achieved with CD25 microbeads (Miltenyi Biotech), less than 1% CD4posCD25pos T cells persisted after depletion and no CD4posCD25high T cells persisted as assessed by FACS analysis (data not shown).
4.4 DC-T cell coculture in the absence of exogenous antigen
For primary culture, 105 immature DC or LPS-matured DC were added to 106 autologous CD4pos T cells in 48-well plates. For some experiments, CD4pos T cells were first depleted in CD4posCD25pos T cells before addition to DC. For T cells phenotyping by FACS and cytokine secretion determination, cells and supernatants were harvested after6 days; for mRNA content analysis, kinetics experiments were performed. To assess T cell proliferation, the cocultures were performed in 96-well plates at the same ratio, i.e. 2×104 DC together with 2×105 T cells. Proliferation was assessed after 6 days by [3H]thymidine (1 μCi/well) incorporation for 16 h. In some experiments, 10 μg/ml neutralizing mAb against HLA-DR, CD80 and CD86 were added at the beginning of the AMLR. For restimulation experiments, T cells were harvested after 6 days of the primary coculture. They were then restimulated with LPS-matured DC generated from fresh blood of the same donor, in the same condition as for the primary stimulation.
4.5 Suppressive T cell function
At the end of the primary AMLR, T cells were harvested and tested for their suppressive function. Graded numbers of primed T lymphocytes were added to an allogeneic MLR consisting in 2×105 CD4+ T lymphocytes from an unrelated donor simulated by 2×104 LPS-matured DC in 96-well plate. For all the experiments, unless specified, DC used as stimulators forthe allogeneic MLR were generated from the same donor as the one used for the primary AMLR. In some experiments, rIL-2 (100 U/ml), rIL-7 (10 ng/ml), rIL-15 (10 ng/ml) were added alone or in conjunction to the coculture. In other experiments, z-vad (10 μM) or 10 μg/ml neutralizing mAb against IL-10, TGF-β or CTLA-4 were added. To assess the effects of supernatants, T cells were primed by mature autologous DC and restimulated during 48 h by mature autologous DC. Supernatants were harvested and added at the ratio 1/1 (v/v) to the suppression assay. T cell proliferation was measured after 5 days of culture at 37°C and an additional 16 h incubation with [3H]thymidine (1 μCi/well) excepted for kinetics experiments. IL-2 production was determined in the supernatant after 48 h and IFN-γ and IL-5 after 5 days.
4.6 Cytokine assays
IFN-γ, IL-10 and IL-2 were measured in the cell culture supernatant by two-site sandwich ELISA systems using mAb from Biosource (detection limit at 15 pg/ml; Fleurus, Belgium). IL-5 levelswere measured by sandwich ELISA using mAb from PharMingen (detection limit at 30 pg/ml; Erembodegem, Belgium).
4.7 Flow cytometry analysis
For immunofluorescence staining, cells were washed and stained for 20 min at 4°C with optimal dilution of each mAb in PBS supplemented with 0.5% BSA and 5% human serum. Cells were washed againand fluorescence intensity staining was analyzed using a FACScan flow-cytometer (FACSCalibur, Becton Dickinson).
4.8 Real-time RT-PCR
mRNA content of 5×105 viable T cells was isolated using the MagNA Pure mRNA extraction kit (Roche Applied Science, Brussels, Belgium) on the MagNA Pure instrument (Roche Applied Science) following manufacturer's instructions ("mRNA I cells" Roche's protocol, final elution volume 100 μl). Cytokine mRNA levels were then quantified by real-time RT-PCR on a Lightcycler instrument (Roche Applied Science) as described 46. Primers and probes were synthesized at Biosource Europe (Nivelles, Belgium) with the following sequences, respectively for IL-2 and TGF-β: CTCACCAGGATgCTCACATTTA and CAGCAACAATTCCTGGCGATA (forward primers), TCCAGAGGTTTGAGTTCTTCTTCT and AAGGCGAAAGCCCTCAATTT (reverse primers), 6Fam-TGCCCAAgAAGGCCACAGAACTG-Tamra-p and 6Fam-CTGCTGGCACCCAGCGACTCG-Tamra-p (Taqman probes). The sequences used for IL-5, IL-10 and IFN-γ have been previously described 46. Cytokine mRNA levels were determined according to a standard curve generated from serial dilutions of a purified DNA, allowing expression of the results in mRNA copies per 5×105 T cells. For FOX-P3 mRNA expression analysis, the following primers were used: CCCACTTACAGGCACTCCTC (forward) and CTTCTCCTTCTCCAGCACCA (reverse). Amplification was carried out for 42 cycles of 20 s at 95°, 20 s at 58°C and 30 s at 72°C. PCR products were detected on a 2% agarose gel using ethidium bromide. As control, mRNA content for β-actin was analyzed in the same sample using the following primers: GGTCTCAAACATGATCTGGG (reverse) and GGGTCAGAAGGATTCCTATG (forward).
4.9 Statistical analysis
Data were compared using the non-parametric Wilcoxon's paired test.
Valérie Verhasselt was supported by the Fonds National de la Recherche Scientifique (Belgium) as "Chargé de Recherches". This work was also supported by an Interuniversity Attraction Pole of the Belgian Federal Government and a Research Concerted Action of the Communauté Française de Belgique. We thank Michel Braun, Alain Le Moine and Oberdan Leo for fruitful discussion and critical reading of this manuscript and Stanislas Goriely for PCR expertise.