Simultaneous Induction of CD4 T Cell Tolerance and CD8 T Cell Immunity by Semimature Dendritic Cells
http://www.100md.com
免疫学杂志 2005年第7期
Abstract
Previous studies suggested that depending on their maturation state, dendritic cells (DC) could either induce T cell tolerance (immature and semimature DC) or T cell activation (mature DC). Pretreatment of C57BL/6 mice with encephalitogenic myelin oligodendrocyte glycoprotein (MOG)35–55 peptide-loaded semimature DC protected from MOG-induced autoimmune encephalomyelitis. This protection was mediated by IL-10-producing CD4 T cells specific for the self Ag. Here we show that semimature DC loaded with the MHC class II-restricted nonself peptide Ag (OVA) induce an identical regulatory T cell cytokine pattern. However, semimature DC loaded simultaneously with MHC class II- and MHC class I-restricted peptides, could efficiently initiate CD8 T cell responses leading to autoimmune diabetes in a TCR-transgenic adoptive transfer model. Double-peptide-loaded semimature DC also induced simultaneously in the same animal partially activated CD8 T cells with cytolytic function as well as protection from MOG-induced autoimmune encephalomyelitis. Our study suggests that the decision between tolerance and immunity not only depends on the DC, but also on the type and activation requirements of the responding T cell.
Introduction
The unique ability of dendritic cells (DC)5 to efficiently activate naive T cells correlates with their property to perform different functions at distinct sites. Immature DC in peripheral tissues respond to inflammatory stimuli and microbial products, mature subsequently and home to lymphoid organs. In contrast to immature peripheral DC, mature DC in lymphoid tissues express high levels of MHC class II (MHC II) and costimulatory molecules, allowing efficient priming of naive T cells (1). Many lines of evidence indicate that, in contrast to their mature counterparts, immature DC can tolerize T cells; this was demonstrated by direct DC-targeting of Ag via DEC-205 (2), cross-presentation induced deletion of Ag-specific CD8+ T cells (3), as well as inducible DC-specific expression of Ag in vivo (4).
However, analysis of spleen sections showed that the majority of DC of naive mice are of an immature phenotype and can still be driven into further maturation (3, 5, 6). This raised the question about roles of these immature DC in lymphoid organs and fostered the attractive hypothesis of immature DC guarantying maintenance of tolerance to peripheral self-Ags (1, 5, 7). Data showing acquisition, processing, and presentation of parietal cell- (8) or pancreatic -cell-derived self-Ags (9) support this possibility. However, other studies showed that "being immature" was not sufficient for DC to exert tolerogenic functions, because only phenotypically mature DC could induce cross-tolerance of CD8+ T cells (10), initiate IL-10-mediated tolerance in a murine model of allergic asthma (11), or induce regulatory T cells, which suppress diabetes in NOD mice (12). Expression of the costimulatory molecules CD80 and/or CD86 on mature DC was also a prerequisite for optimal expansion of suppressive CD4+CD25+ T cells (13). Furthermore, injection of semimature, but not completely immature DC loaded with encephalitogenic myelin oligodendrocyte glycoprotein (MOG) self-peptide induced IL-10-producing CD4+ T cells and protected mice from subsequent induction of experimental autoimmune encephalomyelitis (EAE) with the same peptide (14). This led to a new concept suggesting that semimature DC—the in vitro-created counterparts of naturally occurring steady-state DC in lymphoid organs—can tolerize naive CD4 T cells (15).
In the present study, we investigated whether this concept is applicable also 1) to CD4 T cells specific for foreign Ag and 2) to tolerize CD8 T cells directly or 3) to tolerize CD8 T cells via IL-10-producing bystander CD4 T cells. Here we demonstrate that the same semimature DC loaded with several different antigenic peptides simultaneously induce MHC II-restricted IL-10-producing T cells responsible for protective tolerance to MHC II-restricted Ag as well as immunity toward class I-restricted Ag. Our data indicate that the same semimature DC can induce simultaneously tolerance and immunity, depending on the type of T cell responding to the presented Ag.
Materials and Methods
Mice
Rip-OVAlow (16), OT-I (17), and OT-II (18) mice have been described previously and were kindly provided by C. Kurts (Institute of Molecular Medicine, University Clinics, Bonn, Germany). All mice were bred and maintained at the Institute for Immunology, Ludwig-Maximilians-Universit?t München (Munich, Germany) and were used between 6 and 10 wk of age. All donors and recipients were sex-matched.
Adoptive transfer
CD8+ cells from lymph nodes and spleen of OT-I or OT-II mice were prepared as single cell suspensions. Spleen RBC were removed using ACK buffer (0.15 M NH4Cl, 1 mM KHCO3, 0.1 mM Na2EDTA, pH 7.4) for 4 min at room temperature. The percentage of OT-I or OT-II TCR transgenic T cells among CD8+ or CD4+ T cells, respectively, was controlled via FACS analysis with anti-TCR V5.1/5.2 and V2-specific mAbs. Before injection, cells were washed in PBS and filtered through nylon mesh (Reichelt Chemie Technik; pore size 51 mm). For adoptive T cell transfer, 1 x 106 OT-I and 2.8 x 106 OT-II T cells were injected i.v. into the lateral tail vein of recipient mice.
Dendritic cells
DC were generated from C57BL/6 bone marrow progenitors using GM-CSF as described previously (19). On day 8 of culture, DC were tested for purity by staining with anti-CD11c and anti-MHC II. Purity of DC was routinely around 80%. DC were pulsed for 4 h in complete medium (RPMI 1640 supplemented with 5% FCS, 50 mM 2-ME, and 1% penicillin/strepomycin) with 20 μg/ml OVA257–264, 10 μM OVA323–339, 10 μM MOG35–55 (Neosystem) or mixtures of these. TNF- (500 U/ml) and LPS (1 μg/ml) were added for 4 h while DC were loaded with peptide. The cells were then washed extensively in PBS, and 2.5–4 x 106 DC were injected i.v. into the lateral tail vein of recipient mice.
T cell restimulation and ELISA
Supernatants from spleen cell restimulations were collected after 96 h, and cytokines produced from spleen cells were detected using ELISA kits for IL-10 and IFN- (BD Pharmingen).
Immunostaining and flow cytometry
mAbs used in this study were anti-V5.1/5.2-FITC (MR9-4), anti-V2-biotin (B20.1), anti-CD8-PerCP (53-6.7), and streptavidin-PerCP. All Abs were purchased from BD Pharmingen. A FACSCalibur flow cytometer and CellQuest Software from BD Biosciences were used to collect and analyze the data. Nonviable cells were excluded using forward and side scatter electronic gating.
Induction of EAE and diabetes
MOG35–55 peptide (Biotrend Chemikalien) was used to induce EAE in C57BL/6 mice as described before (14). Briefly, C57BL/6 mice were injected s.c. with 50 μg of MOG35–55 peptide in 50 μl of PBS emulsified in 50 μl of CFA that was further enriched with 10 mg/ml Mycobacterium tuberculosis (H37Ra; Difco/BD Pharmingen). In addition 200 ng of Pertussis toxin (List/Quadratech) were injected i.p. at days 0 and 2. Mice were observed daily for clinical signs of disease. Mice were scored according to their clinical severity of disease as follows: grade 0, no abnormality; grade 1, limp tail; grade 2, limp tail and hind limb weakness (waddling gait); grade 3, partial hind limb paralysis; grade 4, complete hind limb paralysis; and grade 5, moribund. The data are plotted as the mean daily clinical score for all animals in a particular treatment group.
Glucose levels in the urine were checked every day by glucose test strips (Diabur-Test 5000) from Roche Diagnostics. Mice were considered diabetic when glucose concentration was >5.5 mmol/L.
In vivo CTL assay
This assay was performed as described before (20). Syngeneic C57BL/6 spleen and lymph node cells were depleted of erythrocytes by osmotic lysis. Cells were washed and split into two populations. One population was pulsed with 10–6 M OVA257–264-peptide (or in some experiments OVA257–264 plus OVA323–339) for 1 h at 37°C, washed, and labeled with a high concentration of CFSE (2.5 μM; CFSEhigh cells). The second control population was labeled with a low concentration of CFSE (0.25 μM; CFSElow cells). For i.v. injection, an equal number of cells from each population (CFSEhigh and CFSElow) was mixed, such that each mouse received a total of 20 x 106 cells. Cells were injected into mice immunized 8 days before with peptide-pulsed DC. Twenty hours later, mice were sacrificed and spleen and lymph nodes were removed. Cell suspensions were analyzed by flow cytometry; 5 x 105 CFSE-positive cells were collected for analysis. Peptide-pulsed and unpulsed target cells were recognized according to their different CFSE intensities. To calculate specific lysis, the following formulas were used: ratio = (percentage CFSElow/percentage CFSEhigh); percentage specific lysis = (1 – (ratio unprimed/ratio primed) x 100).
Results
Semimature DC can induce IL-10-producing CD4 T cells specific for self and foreign Ag
Semimature DC are indistinguishable from fully mature DC according to surface expression of typical DC maturation markers. However, they lack the ability to produce cytokines and, therefore, induce IL-10-mediated regulatory responses rather than IFN--producing T effector cells. To investigate tolerance induction by TNF--treated semimature DC, we incubated DC from GM-CSF bone-marrow cultures with either LPS or TNF- as described previously (14). Both, LPS and TNF--treated DC were similarly mature according to surface expression of MHC II, CD80, CD86, CD25, MHC class I (MHC I), and CD40 (data not shown). However, as we reported previously (14), in contrast to LPS-DC, semimature TNF-DC did not produce IL-6, IL-12p40, or IL-12p70 (data not shown) and were able to prevent EAE via an IL-10-dependent mechanism, when loaded with MOG35–55-peptide and injected into C57BL/6 mice (Ref. 14 , data not shown, and see Fig. 3).
FIGURE 3. Semimature DC simultaneously tolerize MOG-specific CD4+ T cells and activate OVA-specific CD8+ T cells. C57BL/6 mice were immunized on days –6, –3, and 0 with mature (A, filled bars) or semimature DC (A, hatched bars), loaded with the peptides indicated below, or no DC (A, open bars). On day 8 after the last immunization, four mice of each group were injected with a 1:1 mixture of CFSElow labeled unloaded spleen cells and CFSEhigh labeled spleen cells loaded with the same peptide combinations. After an additional 20 h, mice were sacrificed and spleen cells were analyzed for CFSE-positive cells by flow cytometry ("in vivo killer assay"). A representative example from a FACS-analysis of nonimmunized (left panel) and immunized (right panel) animals from such a killer assay, gated on CFSE+ spleen cells is shown (A, upper panel). The FACS results are displayed as percent specific lysis (A, lower panel). The specific lysis was calculated as described in Materials and Methods. B, On day 3 postimmunization, spleen cells of another four mice from each group were restimulated with the respective peptides in vitro as shown for OVA257–264. Ninety-six hours later supernatants were taken and analyzed by ELISA for IFN- (B, left) and IL-10 (B, right) production. Symbols represent data derived from single mice. C, Another four to six mice from each group, which received MOG-loaded DC were used to induce EAE on day 3 after the last immunization. Animals were observed for paralysis and mean disease scores were determined as described in Materials and Methods. C, The kinetic of EAE development after immunization with either MOG35–55-loaded (left panel) or MOG35–55 + OVA257–264-pulsed (right panel) differently matured DC is shown.
To further assess the immunostimulatory capacities of the differentially matured DC, we immunized mice three times with LPS- or TNF-DC loaded with either the class II-restricted MOG35–55 or OVA323–339 peptide. Restimulation of T cells with the relevant peptide showed in general weaker proliferative T cell responses in those mice injected with TNF-DC as compared with fully mature LPS-DC (Fig. 1A). Also the produced cytokine-patterns were different: while T cells from LPS-DC immunized mice produced both IFN- (Fig. 1B) as well as IL-10 (Fig. 1C) in response to antigenic peptide, the TNF-DC pretreatment lead to strongly diminished IFN- production (Fig. 1B). Therefore, TNF-DC induced Ag-specific IL-10-producing T cells specific for MOG self-Ag (as described previously (14)), but also for the foreign model-Ag OVA.
FIGURE 1. OVA323–339 -pulsed TNF-DC induce identical peptide-specific CD4+ T cell response as TNF-DC pulsed with MOG35–55. C57BL/6 mice received three i.v. injections (days –7, –5, –3) of TNF-DC or LPS/CD40-DC pulsed either with OVA323–339 or MOG35–55 peptide. Splenocytes from these mice were restimulated at day 0 with 2 μM of the respective peptide (; OVA, left graphs or MOG, right graphs) or without peptide (). T cell responses were assessed after 72 h of culture. B and C, Cell supernatants were harvested and their cytokine content measured by ELISA. A, Proliferation was assayed by [3H]thymidine incorporation for an additional 18 h.
Semimature DC elicit diabetes in RIP-OVAlow mice
To analyze if semimature DC might be used to either directly or indirectly tolerize CD8+ T cells and have therapeutic effects in CD8+ T cell-mediated autoimmune diseases, we usedthe OT-I adoptive transfer model to induce diabetes in RIP-OVAlow mice. Transgenic RIP-OVAlow mice, which express OVA as model autoantigen under control of the rat insulin promoter, are tolerant to OVA, which is expressed on the cells of the pancreatic islets (16). Upon transfer into these hosts, naive "autoreactive" OVA-specific CD8+ CTL from TCR-transgenic OT-I mice, are ignorant to their specific Ag due to low amounts of OVA expression in RIP-OVAlow recipients (16, 21, 22, 23). However, when OT-I T cells become activated in vivo, they expand and destroy the OVA-expressing pancreas inducing diabetes as measured by high urine glucose levels (16, 21, 22, 23). To examine the tolerizing capacity of semimature DC in this system, OVA257–264-pulsed TNF-DC or LPS-DC were injected into RIP-OVAlow mice reconstituted with 1 x 106 OT-I cells. To monitor T cell expansion, blood was analyzed and the percentage of OT-I T cells determined by flow cytometry (Fig. 2A). After i.v. immunization OT-I T cells expanded to over 70% of CD8+ PBL (Fig. 2A). The kinetic of OT-I expansion induced by mature or semimature DC was similar in all mice, peaked at day 4 postinfection (Fig. 2A) and was comparable, when PBL (Fig. 2A) and spleens (data not shown) were analyzed. However, because T cell expansion not only accompanies activation but also tolerization of T cells (24, 25), we determined diabetes induction as a sign for CTL activation in vivo. Although all mice treated with mature or semimature DC were clinically scored as diabetic (Fig. 2B, left panel), the urine glucose levels in the group treated with TNF-DC was in average 2- to 3-fold lower (Fig. 2B, left panel). In addition, while LPS-DC-treated animals did not clinically improve, the TNF-DC-treated mice fully recovered from diabetes after 27 days (Fig. 2B, left panel). However, after complete recovery of the animals vaccinated with TNF-DC (day 27, Fig. 2B, left panel), a second immunization of the recovered animals with OVA257–264 (100 μg in CFA s.c.) induced strong diabetes with no recovery and, therefore, excluded the possibility that CD8 T cell tolerance induction by TNF-DC could be accompanied by transient acquisition of cytolytic function (Fig. 2C). Taken together, these data suggest that semimature TNF-DC have no direct tolerogenic capacity toward CD8+ T cells, but rather activate them to effector killer cells. However, this activation was weaker as compared with CTL induced by fully mature LPS-DC.
FIGURE 2. Semimature DC induce cytotoxicity in transgenic CD8+ OT-I T cells. A, RIP-OVAlow mice received 1 x 106 CD8+ OT-I T cells i.v. and were immunized 1 day later with 3 x 106 either LPS- or TNF--matured, OVA257–264-loaded DC or not immunized. A, Representative flow cytometry dot plots of blood analysis from day 4 after immunization. Cells within the lymphocyte gate of forward scatter/side scatter analysis (data not shown) were first analyzed for expression of CD8 (data not shown). CD8+ lymphocytes were further analyzed for expression of the transgenic TCR-V segments 2 and 5.1, expressed by OT-I T cells (A, dot plots). The mean percentage of OT-I T cells within CD8+ peripheral blood lymphocytes and the SDs are indicated for every treatment group (n = 6). The kinetics of OT-I T cell expansion after immunization with the different DC is shown in the right graph. Mice from this experiment were also monitored for induction of diabetes by measuring glucosuria. B, Upper left panel, course of diabetes development in Rip-OVAlow mice after immunization with differently matured, OVA257–264-loaded DC. In the lower left panel, the same data are displayed as percent of diabetic mice. Right panel, In this experiment RIP-OVAlow mice received both 1 x 106 (CD8+) OT-I and 2.8 x 106 (CD4+) OT-II cells and DC for immunization were pulsed with both OVA257–264 (MHC I) and OVA323–339 (MHC II). In the upper graph, the course of diabetes development after immunization is shown; the lower graph displays the same data as percent of diabetic mice (n = 4 mice per group). C, Mice from TNF-DC and control groups from the experiment shown in (B, left panel) without signs of diabetes at day 27 after DC-vaccination were immunized with 100 μg of OVA257–264 (MHC I)-peptide emulsified in CFA s.c. to determine whether remaining OT-I cells were tolerant. The course of diabetes development was monitored.
It has previously been shown, that TNF-DC can mediate tolerance via IL-10-producing Ag-specific CD4 regulatory T cells (14). Because RIP-OVAlow mice are tolerant for OVA and their repertoire of T cells specific for OVA is either purged or tolerized (16, 21, 22), we studied whether provision of an exogenous source of OVA-specific CD4 Th cells from TCR-transgenic OT-II mice might overcome the absence of tolerance induction by TNF-DC.
Therefore, we cotransferred both OT-I and OT-II cells into RIP-OVAlow mice and treated them with LPS-DC or TNF-DC pulsed simultaneously with both OVA257–264 and OVA323–339. However, also in the presence of CD4+ OT-II cells and their specific peptide on DC, mice from both experimental groups developed diabetes (Fig. 2B, right panel). Moreover, mice immunized with OVA257–264/OVA323–339-loaded TNF-DC did not recover from diabetes, and their urine glucose levels were similarly elevated as compared with mice immunized with fully mature DC (Fig. 2B, right panel). So far our data indicate that TNF-DC are neither capable without (Fig. 2B) or with (Fig. 2B) the help of Ag-specific CD4 cells to tolerize OVA-specific OT-I CD8 T cells and rather activate CD8 cells to become cytotoxic effector cells.
Semimature DC induce CTL function in C57BL/6 mice and simultaneously protect from CD4 T cell-mediated EAE
To exclude that absence of tolerance induction by TNF-DC was due to the transgenic nature of OT-I or OT-II T cells, or their artificially elevated precursor frequencies, we next compared polyclonal CTL responses induced by mature and semimature DC. To analyze the cytotoxic activity induced by the different DC, we performed "in vivo killer assays" (20) in normal C57BL/6 mice. DC-treated mice were injected with a 1:1 mixture of differently CFSE-labeled spleen cell targets, either pulsed with OVA257–264-peptide (CFSE high) or unpulsed (CFSE low) (Fig. 3A). CTLs generated by immunization of the mice with mature OVA257–264-loaded DC, selectively killed the peptide-loaded (CFSE high) target cell population (Fig. 3A) within a few hours, allowing calculation of specific lysis (20). As shown in Fig. 3A, killing of injected OVA257–264-pulsed spleen cells was not statistically different in mice treated with OVA257–264-loaded LPS- or TNF-DC. Furthermore, when CD4 T cell epitopes of MOG or OVA or both were provided together with the OVA257–264 MHC I epitope on DC as well as on the targets, similar specific killing was observed in each group (Fig. 3A). These results indicate, that neither OVA- nor MOG-specific IL-10-producing CD4 T cells (Fig. 1) suppressed the CD8 T cell responses as bystanders.
We next analyzed cytokine production of CD8 T cells induced by mature or semimature DC. Restimulated splenocytes from mice treated with OVA257–264-pulsed LPS-DC produced mainly IFN-, but no or little IL-10, while semimature TNF-DC did not elicit any significant cytokine responses at all (Fig. 3B). The absence of IFN- secretion by CD8 T cells from mice immunized with TNF-DC (Fig. 3B) in contrast to those treated with fully mature DC indicates incomplete induction of CTL responses in vivo. However, these IFN- negative CD8 T cells do show cytolytic functions (Figs. 2B and 3A).
To proof that TNF-DC loaded with either MOG35–55 or both OVA257–264 and MOG35–55 were similarly tolerogenic for MOG-specific T cells and able to protect mice from EAE, we induced EAE in mice treated as in Fig. 3A. As published before (14), immunization with semimature MOG35–55 -loaded DC protected C57BL/6 mice from EAE, while mice without DC-treatment or injected with mature MOG35–55-pulsed LPS-DC developed severe EAE (Fig. 3C, MOG). However, the same semimature DC loaded with both, MOG35–55 and OVA257–264, that induced specific CTLs against the MHC I peptide OVA257–264 (Fig. 3A) protected mice from MOG-induced EAE (Fig. 3C). These data suggest that the same semimature DC, which induce MOG-specific IL-10-producing CD4 (regulatory) T cells triggers simultaneously cytotoxic effector T cells, which kill targets (Fig. 3A) in absence of IFN- production. The latter were not inhibited by bystander regulatory T cells triggered most likely in their vicinity, because efficient cytolytic activity was induced also by double-, and triple-loaded semimature DC (Fig. 3B).
Taken together, these data suggest that the tolerizing capacity of semimature DC is restricted to Ag-specific CD4 T cells, but ineffective to suppress TCR transgenic (Fig. 2) as well as polyclonal CD8 T cells (Fig. 3) specific for OVA.
Discussion
In vivo counterparts to the in vitro-generated TNF-DC used in this study could be DC in the lymphatics, which migrate from nonlymphoid organs to peripheral lymph nodes, transporting apoptotic cells (26) or tissue (self-)Ag (8, 9, 27). In the lymph node and spleen, these migrated DC are crucial to induce peripheral tolerance (2, 28, 29). The semimature stage of these "steady-state" DC might be attributed to lack of activation by pathogens via TLR ligation leading to phenotypically mature DC that do not produce IL-12 (data not shown and Ref. 14). However, also active suppression of IL-12 production in mature DC has been described for helminth Ag, such as schistosome egg Ag (30) and induction of a general block of DC-maturation was demonstrated for secreted Ag from M. tuberculosis (31). Therefore, several arguments emerged, which support the existence of semimature DC in vivo induced either by lack of appropriate signals or pathogen-induced maturation blockade. It is therefore important to understand the functions of such semimature DC in more detail.
In vitro-cultured TNF--treated semimature DC loaded with MOG35–55 peptide induce MOG35–55-specific tolerance and protect C57BL/6 mice from MOG35–55-induced EAE (Ref. 14 and Fig. 3C). As experiments with semimature DC from IL-10–/– mice showed, DC do not have to produce IL-10 themselves to excert tolerogenic capacity (14). This is in contrast to IL-10-producing DC protecting from EAE (32, 33) or inducing mucosal tolerance (11) in other studies. Similar to immature human DC (34, 35) the mechanism of MOG-tolerance induction by semimature DC was revealed to be dependent on IL-10-producing regulatory T cells (14), a mechanism that has been demonstrated to act also through bystander suppression (36, 37).
Therefore, we investigated the tolerizing capacity of semimature DC to CD4 T cells specific for nonself Ag and CD8+ T cells in addition to the described MOG-self Ag specific CD4 T cell response (14). We show that semimature DC induce a similar IL-10-biased cytokine pattern in foreign Ag (OVA)-specific CD4 T cells as observed for self-Ag-specific MOG T cells (Fig. 1).
When adoptively transferred transgenic OT-I or endogenous polyclonal CD8+ T cells were exposed to MHC I specific peptide-loaded semimature DC in vivo, T cell expansion (Fig. 2) as well as cytotoxic T cell effector functions (Figs. 2 and 3) were induced. Also the addition of exogenous IL-10-producing CD4 T cells (Fig. 2) and simultaneous presentation of corresponding CD4 peptides by semimature DC (Figs. 2 and 3) did not lead to tolerance induction, but rather enhanced CD8 T cell responses (Fig. 2). The magnitude of CD8 responses (Fig. 2) was lower and CD8 cytokine production (Fig. 3) absent, when the responses were initiated by semimature DC. This phenomenon could 1) be due to absent IL-12 production by DC (Ref. 14 and data not shown) and 2) be overcome by addition of exogenous Ag-specific CD4+ T cell help (Fig. 2). Similar to another report (10), where presence of CD4+ T cells altered tolerogenic DC in activating DC, we detected an enhancement of CD8 T cell immunity (Fig. 2). When DC were simultaneously loaded with the tolerogenic MOG35–55 peptide as well as OVA peptides, the same DC induced tolerance for MOG35–55 (Fig. 3C), but immunity for OVA by triggering OVA-specific CD8+ cytotoxic T cell responses (Figs. 2B and 3A). This latter experiment indicated that IL-10-producing MOG- or OVA-peptide specific regulatory CD4 T cells were insufficient to exert bystander suppression on CD8 T cells, neither during priming by the same double- or triple (MOG/OVA)-loaded DC, nor at the level of recognition of (MOG/OVA)-presenting targets (Fig. 3). Data from other experimental systems have suggested that bystander suppression is mediated through IL-10, TGF- or cell-cell contacts (36, 37). Hypothetically the semimature DC induce or expand IL-10-producing MOG-specific regulatory T cells, which then may suppress autoreactive clones of similar specificities at other sites and at later time points (for example, when the disease is induced). However, it is not clear yet whether CD8 T cells have similar requirements for tolerization like MOG-specific EAE-inducing CD4 T cells. Therefore, it is possible that semimature DC cannot tolerize CD8 T cells efficiently, suggesting differential requirements for CD4 and CD8 T cells. Another potential reason for the observed differences between CD4 and CD8 T cells might be the differential loading efficiency of TNF-DC with MHC I vs II peptides. In general, much more peptide is required to trigger the effector function of CD4 T cells when compared with CD8 T cells, and this could also account for the phenotype described in this study. Although we tried in our experiments to work under saturating conditions for peptide loading, an adequate comparison of MHC I- and II-induced responses seems to be impossible.
The OVA-specific CTL induced by TNF-DC were unable to produce IFN-, which is an indication that their activation was not complete (Fig. 3B). Therefore, together with the data from Fig. 2, one could speculate that a reason for this observed cytotoxic activity could be the relatively low threshold signals needed to trigger CTL-activity: one or few specific TCR/peptide-MHC interactions are necessary to induce the cytolytic machinery (38, 39) and even partially activated CD8 T cells can gain cytoloytic activity (40, 41). In earlier studies the phenomenon of cytotoxic activity in absence of cytokine production has been called "split anergy" (42, 43). Taken together, "CTL" induced by IL-12– TNF-DC might be severely impaired in their tumor- (44) and virus-clearing (45) efficiencies, but are definitely not tolerized.
Overall our results suggest that one DC can exert different functions on CD4 and CD8 T cells simultaneously: induction of CD4 T cell tolerance as well as CD8 T cell immunity. However, it remains an open question, if such an artificially enforced experimental situation of a double-peptide-loaded semimature DC is at all existing in vivo. A potential restriction for such a scenario could be the intrinsically different Ag uptake, processing, and presentation capacities of distinct DC-subsets; i.v. injected soluble OVA has been shown to be taken up by splenic DC, which after isolation, markedly differed in their capacity to present the processed OVA. Although the CD8+ DC predominantly cross-presented OVA to CD8 T cells but showed little stimulation of CD4 T cells, the contrary was observed for the CD8– DC subsets which readily presented OVA in an MHC II-restricted fashion but poorly on MHC I molecules (46).
However, other examples for a segregation of function within one APC have been published before. It was shown that DC copulsed with microbial and helminth Ag were able to prime Th1 responses for the microbial Ag as well as Th2 responses for the helminth Ag (47). Therefore, there is increasing evidence for differential handling of different Ag within the same APC. As a consequence, subsequent induction of qualitatively different immune responses by one and the same DC could appear.
Specifically relevant for the question of maintenance of tolerance to apoptotic self-material were recent findings demonstrating that apoptotic (self-) material and bacterial Ag are handled by discrete intracellular phagosomal maturation and processing pathways in APC, depending on their respective TLR signaling (48). Such a signaling by bacterial material induced phagocytosis and processing pathways, which were independently regulated from constitutive pathways for apoptotic material (48). It seems an attractive hypothesis that DC are equipped with endogenous control mechanisms, allowing them to integrate processing and presentation requirements for different Ags to avoid in vivo situations like the one induced by unphysiological external peptide loading in our study.
Although the physiological occurrence of such a simultaneous induction of tolerance and immunity in vivo seems unlikely, our findings have implications for in vitro-loaded DC as they are currently used for human cancer immunotherapy. When patient-derived DC are loaded with peptides on MHC I and/or II molecules but are improperly matured, the possibility of simultaneous CD4 tolerance and only partial CD8 immunity should be taken into consideration. In addition, these data indicate that CD4 and CD8 T cells have different activation requirements from the side of the APC. Althoughthe semimature state of the presenting DC is suboptimal and therefore still tolerogenic for CD4 T cells, the costimulatory and cytokine equipment is sufficient to induce CTL activity, however, with reduced/absent IFN- production.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank A. Bol and W. Mertl for expert animal facility maintenance.
Footnotes
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.
1 This work was supported by the Deutsche Forschungsgemeinschaft Grant Br 1889/2, SFB571, SFB456 and Wilhelm-Sander Stiftung for T.B.; and Grant LU 851/2-1 and the SFB643 (to M.B.L.); and the ELAN program of the University Hospital Erlangen (to C.W.).
2 P.K. and C.W. contributed equally.
3 M.B.L. and T.B. contributed equally.
4 Address correspondence and reprint requests to Dr. Thomas Brocker, Institute for Immunology, Ludwig-Maximilians-Universit?t München, Goethestrasse 31, 80336 München, Germany. E-mail address: tbrocker@ifi.med.uni-muenchen.de
5 Abbreviations used in this paper: DC, dendritic cell; MOG, myelin oligodendrocyte glycoprotein; MHC II, MHC class II; MHC I, MHC class I; EAE, experimental autoimmune encephalomyelitis.
Received for publication September 7, 2004. Accepted for publication December 6, 2004.
References
Banchereau, J., R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392:245.
Hawiger, D., K. Inaba, Y. Dorsett, M. Guo, K. Mahnke, M. Rivera, J. V. Ravetch, R. M. Steinman, M. C. Nussenzweig. 2001. Dendritic cells induce peripheral T cell unresponsiveness under steady state conditions in vivo. J. Exp. Med. 194:769.
Bonifaz, L., D. Bonnyay, K. Mahnke, M. Rivera, M. C. Nussenzweig, R. M. Steinman. 2002. Efficient targeting of protein antigen to the dendritic cell receptor DEC-205 in the steady state leads to antigen presentation on major histocompatibility complex class I products and peripheral CD8+ T cell tolerance. J. Exp. Med. 196:1627.
Probst, H. C., J. Lagnel, G. Kollias, M. van den Broek. 2003. Inducible transgenic mice reveal resting dendritic cells as potent inducers of CD8+ T cell tolerance. Immunity 18:713.
Wilson, N. S., D. El-Sukkari, G. T. Belz, C. M. Smith, R. J. Steptoe, W. R. Heath, K. Shortman, J. A. Villadangos. 2003. Most lymphoid organ dendritic cell types are phenotypically and functionally immature. Blood 102:2187.
De Smedt, T., B. Pajak, E. Muraille, L. Lespagnard, E. Heinen, P. De Baetselier, J. Urbain, O. Leo, M. Moser. 1996. Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J. Exp. Med. 184:1413.
Steinman, R. M., M. C. Nussenzweig. 2002. Avoiding horror autotoxicus: the importance of dendritic cells in peripheral T cell tolerance. Proc. Natl. Acad. Sci. USA 99:351.
Scheinecker, C., R. McHugh, E. M. Shevach, R. N. Germain. 2002. Constitutive presentation of a natural tissue autoantigen exclusively by dendritic cells in the draining lymph node. J. Exp. Med. 196:1079.
Turley, S., L. Poirot, M. Hattori, C. Benoist, D. Mathis. 2003. Physiological cell death triggers priming of self-reactive T cells by dendritic cells in a type-1 diabetes model. J. Exp. Med. 198:1527.
Albert, M. L., M. Jegathesan, R. B. Darnell. 2001. Dendritic cell maturation is required for the cross-tolerization of CD8+ T cells. Nat. Immunol. 2:1010
Akbari, O., R. H. DeKruyff, D. T. Umetsu. 2001. Pulmonary dendritic cells producing IL-10 mediate tolerance induced by respiratory exposure to antigen. Nat. Immunol. 2:725.
Salomon, B., D. J. Lenschow, L. Rhee, N. Ashourian, B. Singh, A. Sharpe, J. A. Bluestone. 2000. B7/CD28 costimulation is essential for the homeostasis of the CD4+CD25+ immunoregulatory T cells that control autoimmune diabetes. Immunity 12:431
Yamazaki, S., T. Iyoda, K. Tarbell, K. Olson, K. Velinzon, K. Inaba, R. M. Steinman. 2003. Direct expansion of functional CD25+ CD4+ regulatory T cells by antigen-processing dendritic cells. J. Exp. Med. 198:235.
Menges, M., S. Rossner, C. Voigtlander, H. Schindler, N. A. Kukutsch, C. Bogdan, K. Erb, G. Schuler, M. B. Lutz. 2002. Repetitive injections of dendritic cells matured with tumor necrosis factor induce antigen-specific protection of mice from autoimmunity. J. Exp. Med. 195:15.
Lutz, M. B., G. Schuler. 2002. Immature, semi-mature and fully mature dendritic cells: which signals induce tolerance or immunity?. Trends Immunol. 23:445.
Blanas, E., F. R. Carbone, J. Allison, J. F. Miller, W. R. Heath. 1996. Induction of autoimmune diabetes by oral administration of autoantigen. Science 274:1707.
Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, F. R. Carbone. 1994. T cell receptor antagonist peptides induce positive selection. Cell 76:17.
Barnden, M. J., J. Allison, W. R. Heath, F. R. Carbone. 1998. Defective TCR expression in transgenic mice constructed using cDNA-based - and -chain genes under the control of heterologous regulatory elements. Immunol. Cell Biol. 76:34.
Lutz, M. B., N. Kukutsch, A. L. Ogilvie, S. Rossner, F. Koch, N. Romani, G. Schuler. 1999. An advanced culture method for generating large quantities of highly pure dendritic cells from mouse bone marrow. J. Immunol. Methods 223:77.
Coles, R. M., S. N. Mueller, W. R. Heath, F. R. Carbone, A. G. Brooks. 2002. Progression of armed CTL from draining lymph node to spleen shortly after localized infection with herpes simplex virus 1. J. Immunol. 168:834.
Kurts, C., R. M. Sutherland, G. Davey, M. Li, A. M. Lew, E. Blanas, F. R. Carbone, J. F. Miller, W. R. Heath. 1999. CD8 T cell ignorance or tolerance to islet antigens depends on antigen dose. Proc. Natl. Acad. Sci. USA 96:12703.
Kurts, C., J. F. Miller, R. M. Subramaniam, F. R. Carbone, W. R. Heath. 1998. Major histocompatibility complex class I-restricted cross-presentation is biased towards high dose antigens and those released during cellular destruction. J. Exp. Med. 188:409.(Petra Kleindienst, Carste)
Previous studies suggested that depending on their maturation state, dendritic cells (DC) could either induce T cell tolerance (immature and semimature DC) or T cell activation (mature DC). Pretreatment of C57BL/6 mice with encephalitogenic myelin oligodendrocyte glycoprotein (MOG)35–55 peptide-loaded semimature DC protected from MOG-induced autoimmune encephalomyelitis. This protection was mediated by IL-10-producing CD4 T cells specific for the self Ag. Here we show that semimature DC loaded with the MHC class II-restricted nonself peptide Ag (OVA) induce an identical regulatory T cell cytokine pattern. However, semimature DC loaded simultaneously with MHC class II- and MHC class I-restricted peptides, could efficiently initiate CD8 T cell responses leading to autoimmune diabetes in a TCR-transgenic adoptive transfer model. Double-peptide-loaded semimature DC also induced simultaneously in the same animal partially activated CD8 T cells with cytolytic function as well as protection from MOG-induced autoimmune encephalomyelitis. Our study suggests that the decision between tolerance and immunity not only depends on the DC, but also on the type and activation requirements of the responding T cell.
Introduction
The unique ability of dendritic cells (DC)5 to efficiently activate naive T cells correlates with their property to perform different functions at distinct sites. Immature DC in peripheral tissues respond to inflammatory stimuli and microbial products, mature subsequently and home to lymphoid organs. In contrast to immature peripheral DC, mature DC in lymphoid tissues express high levels of MHC class II (MHC II) and costimulatory molecules, allowing efficient priming of naive T cells (1). Many lines of evidence indicate that, in contrast to their mature counterparts, immature DC can tolerize T cells; this was demonstrated by direct DC-targeting of Ag via DEC-205 (2), cross-presentation induced deletion of Ag-specific CD8+ T cells (3), as well as inducible DC-specific expression of Ag in vivo (4).
However, analysis of spleen sections showed that the majority of DC of naive mice are of an immature phenotype and can still be driven into further maturation (3, 5, 6). This raised the question about roles of these immature DC in lymphoid organs and fostered the attractive hypothesis of immature DC guarantying maintenance of tolerance to peripheral self-Ags (1, 5, 7). Data showing acquisition, processing, and presentation of parietal cell- (8) or pancreatic -cell-derived self-Ags (9) support this possibility. However, other studies showed that "being immature" was not sufficient for DC to exert tolerogenic functions, because only phenotypically mature DC could induce cross-tolerance of CD8+ T cells (10), initiate IL-10-mediated tolerance in a murine model of allergic asthma (11), or induce regulatory T cells, which suppress diabetes in NOD mice (12). Expression of the costimulatory molecules CD80 and/or CD86 on mature DC was also a prerequisite for optimal expansion of suppressive CD4+CD25+ T cells (13). Furthermore, injection of semimature, but not completely immature DC loaded with encephalitogenic myelin oligodendrocyte glycoprotein (MOG) self-peptide induced IL-10-producing CD4+ T cells and protected mice from subsequent induction of experimental autoimmune encephalomyelitis (EAE) with the same peptide (14). This led to a new concept suggesting that semimature DC—the in vitro-created counterparts of naturally occurring steady-state DC in lymphoid organs—can tolerize naive CD4 T cells (15).
In the present study, we investigated whether this concept is applicable also 1) to CD4 T cells specific for foreign Ag and 2) to tolerize CD8 T cells directly or 3) to tolerize CD8 T cells via IL-10-producing bystander CD4 T cells. Here we demonstrate that the same semimature DC loaded with several different antigenic peptides simultaneously induce MHC II-restricted IL-10-producing T cells responsible for protective tolerance to MHC II-restricted Ag as well as immunity toward class I-restricted Ag. Our data indicate that the same semimature DC can induce simultaneously tolerance and immunity, depending on the type of T cell responding to the presented Ag.
Materials and Methods
Mice
Rip-OVAlow (16), OT-I (17), and OT-II (18) mice have been described previously and were kindly provided by C. Kurts (Institute of Molecular Medicine, University Clinics, Bonn, Germany). All mice were bred and maintained at the Institute for Immunology, Ludwig-Maximilians-Universit?t München (Munich, Germany) and were used between 6 and 10 wk of age. All donors and recipients were sex-matched.
Adoptive transfer
CD8+ cells from lymph nodes and spleen of OT-I or OT-II mice were prepared as single cell suspensions. Spleen RBC were removed using ACK buffer (0.15 M NH4Cl, 1 mM KHCO3, 0.1 mM Na2EDTA, pH 7.4) for 4 min at room temperature. The percentage of OT-I or OT-II TCR transgenic T cells among CD8+ or CD4+ T cells, respectively, was controlled via FACS analysis with anti-TCR V5.1/5.2 and V2-specific mAbs. Before injection, cells were washed in PBS and filtered through nylon mesh (Reichelt Chemie Technik; pore size 51 mm). For adoptive T cell transfer, 1 x 106 OT-I and 2.8 x 106 OT-II T cells were injected i.v. into the lateral tail vein of recipient mice.
Dendritic cells
DC were generated from C57BL/6 bone marrow progenitors using GM-CSF as described previously (19). On day 8 of culture, DC were tested for purity by staining with anti-CD11c and anti-MHC II. Purity of DC was routinely around 80%. DC were pulsed for 4 h in complete medium (RPMI 1640 supplemented with 5% FCS, 50 mM 2-ME, and 1% penicillin/strepomycin) with 20 μg/ml OVA257–264, 10 μM OVA323–339, 10 μM MOG35–55 (Neosystem) or mixtures of these. TNF- (500 U/ml) and LPS (1 μg/ml) were added for 4 h while DC were loaded with peptide. The cells were then washed extensively in PBS, and 2.5–4 x 106 DC were injected i.v. into the lateral tail vein of recipient mice.
T cell restimulation and ELISA
Supernatants from spleen cell restimulations were collected after 96 h, and cytokines produced from spleen cells were detected using ELISA kits for IL-10 and IFN- (BD Pharmingen).
Immunostaining and flow cytometry
mAbs used in this study were anti-V5.1/5.2-FITC (MR9-4), anti-V2-biotin (B20.1), anti-CD8-PerCP (53-6.7), and streptavidin-PerCP. All Abs were purchased from BD Pharmingen. A FACSCalibur flow cytometer and CellQuest Software from BD Biosciences were used to collect and analyze the data. Nonviable cells were excluded using forward and side scatter electronic gating.
Induction of EAE and diabetes
MOG35–55 peptide (Biotrend Chemikalien) was used to induce EAE in C57BL/6 mice as described before (14). Briefly, C57BL/6 mice were injected s.c. with 50 μg of MOG35–55 peptide in 50 μl of PBS emulsified in 50 μl of CFA that was further enriched with 10 mg/ml Mycobacterium tuberculosis (H37Ra; Difco/BD Pharmingen). In addition 200 ng of Pertussis toxin (List/Quadratech) were injected i.p. at days 0 and 2. Mice were observed daily for clinical signs of disease. Mice were scored according to their clinical severity of disease as follows: grade 0, no abnormality; grade 1, limp tail; grade 2, limp tail and hind limb weakness (waddling gait); grade 3, partial hind limb paralysis; grade 4, complete hind limb paralysis; and grade 5, moribund. The data are plotted as the mean daily clinical score for all animals in a particular treatment group.
Glucose levels in the urine were checked every day by glucose test strips (Diabur-Test 5000) from Roche Diagnostics. Mice were considered diabetic when glucose concentration was >5.5 mmol/L.
In vivo CTL assay
This assay was performed as described before (20). Syngeneic C57BL/6 spleen and lymph node cells were depleted of erythrocytes by osmotic lysis. Cells were washed and split into two populations. One population was pulsed with 10–6 M OVA257–264-peptide (or in some experiments OVA257–264 plus OVA323–339) for 1 h at 37°C, washed, and labeled with a high concentration of CFSE (2.5 μM; CFSEhigh cells). The second control population was labeled with a low concentration of CFSE (0.25 μM; CFSElow cells). For i.v. injection, an equal number of cells from each population (CFSEhigh and CFSElow) was mixed, such that each mouse received a total of 20 x 106 cells. Cells were injected into mice immunized 8 days before with peptide-pulsed DC. Twenty hours later, mice were sacrificed and spleen and lymph nodes were removed. Cell suspensions were analyzed by flow cytometry; 5 x 105 CFSE-positive cells were collected for analysis. Peptide-pulsed and unpulsed target cells were recognized according to their different CFSE intensities. To calculate specific lysis, the following formulas were used: ratio = (percentage CFSElow/percentage CFSEhigh); percentage specific lysis = (1 – (ratio unprimed/ratio primed) x 100).
Results
Semimature DC can induce IL-10-producing CD4 T cells specific for self and foreign Ag
Semimature DC are indistinguishable from fully mature DC according to surface expression of typical DC maturation markers. However, they lack the ability to produce cytokines and, therefore, induce IL-10-mediated regulatory responses rather than IFN--producing T effector cells. To investigate tolerance induction by TNF--treated semimature DC, we incubated DC from GM-CSF bone-marrow cultures with either LPS or TNF- as described previously (14). Both, LPS and TNF--treated DC were similarly mature according to surface expression of MHC II, CD80, CD86, CD25, MHC class I (MHC I), and CD40 (data not shown). However, as we reported previously (14), in contrast to LPS-DC, semimature TNF-DC did not produce IL-6, IL-12p40, or IL-12p70 (data not shown) and were able to prevent EAE via an IL-10-dependent mechanism, when loaded with MOG35–55-peptide and injected into C57BL/6 mice (Ref. 14 , data not shown, and see Fig. 3).
FIGURE 3. Semimature DC simultaneously tolerize MOG-specific CD4+ T cells and activate OVA-specific CD8+ T cells. C57BL/6 mice were immunized on days –6, –3, and 0 with mature (A, filled bars) or semimature DC (A, hatched bars), loaded with the peptides indicated below, or no DC (A, open bars). On day 8 after the last immunization, four mice of each group were injected with a 1:1 mixture of CFSElow labeled unloaded spleen cells and CFSEhigh labeled spleen cells loaded with the same peptide combinations. After an additional 20 h, mice were sacrificed and spleen cells were analyzed for CFSE-positive cells by flow cytometry ("in vivo killer assay"). A representative example from a FACS-analysis of nonimmunized (left panel) and immunized (right panel) animals from such a killer assay, gated on CFSE+ spleen cells is shown (A, upper panel). The FACS results are displayed as percent specific lysis (A, lower panel). The specific lysis was calculated as described in Materials and Methods. B, On day 3 postimmunization, spleen cells of another four mice from each group were restimulated with the respective peptides in vitro as shown for OVA257–264. Ninety-six hours later supernatants were taken and analyzed by ELISA for IFN- (B, left) and IL-10 (B, right) production. Symbols represent data derived from single mice. C, Another four to six mice from each group, which received MOG-loaded DC were used to induce EAE on day 3 after the last immunization. Animals were observed for paralysis and mean disease scores were determined as described in Materials and Methods. C, The kinetic of EAE development after immunization with either MOG35–55-loaded (left panel) or MOG35–55 + OVA257–264-pulsed (right panel) differently matured DC is shown.
To further assess the immunostimulatory capacities of the differentially matured DC, we immunized mice three times with LPS- or TNF-DC loaded with either the class II-restricted MOG35–55 or OVA323–339 peptide. Restimulation of T cells with the relevant peptide showed in general weaker proliferative T cell responses in those mice injected with TNF-DC as compared with fully mature LPS-DC (Fig. 1A). Also the produced cytokine-patterns were different: while T cells from LPS-DC immunized mice produced both IFN- (Fig. 1B) as well as IL-10 (Fig. 1C) in response to antigenic peptide, the TNF-DC pretreatment lead to strongly diminished IFN- production (Fig. 1B). Therefore, TNF-DC induced Ag-specific IL-10-producing T cells specific for MOG self-Ag (as described previously (14)), but also for the foreign model-Ag OVA.
FIGURE 1. OVA323–339 -pulsed TNF-DC induce identical peptide-specific CD4+ T cell response as TNF-DC pulsed with MOG35–55. C57BL/6 mice received three i.v. injections (days –7, –5, –3) of TNF-DC or LPS/CD40-DC pulsed either with OVA323–339 or MOG35–55 peptide. Splenocytes from these mice were restimulated at day 0 with 2 μM of the respective peptide (; OVA, left graphs or MOG, right graphs) or without peptide (). T cell responses were assessed after 72 h of culture. B and C, Cell supernatants were harvested and their cytokine content measured by ELISA. A, Proliferation was assayed by [3H]thymidine incorporation for an additional 18 h.
Semimature DC elicit diabetes in RIP-OVAlow mice
To analyze if semimature DC might be used to either directly or indirectly tolerize CD8+ T cells and have therapeutic effects in CD8+ T cell-mediated autoimmune diseases, we usedthe OT-I adoptive transfer model to induce diabetes in RIP-OVAlow mice. Transgenic RIP-OVAlow mice, which express OVA as model autoantigen under control of the rat insulin promoter, are tolerant to OVA, which is expressed on the cells of the pancreatic islets (16). Upon transfer into these hosts, naive "autoreactive" OVA-specific CD8+ CTL from TCR-transgenic OT-I mice, are ignorant to their specific Ag due to low amounts of OVA expression in RIP-OVAlow recipients (16, 21, 22, 23). However, when OT-I T cells become activated in vivo, they expand and destroy the OVA-expressing pancreas inducing diabetes as measured by high urine glucose levels (16, 21, 22, 23). To examine the tolerizing capacity of semimature DC in this system, OVA257–264-pulsed TNF-DC or LPS-DC were injected into RIP-OVAlow mice reconstituted with 1 x 106 OT-I cells. To monitor T cell expansion, blood was analyzed and the percentage of OT-I T cells determined by flow cytometry (Fig. 2A). After i.v. immunization OT-I T cells expanded to over 70% of CD8+ PBL (Fig. 2A). The kinetic of OT-I expansion induced by mature or semimature DC was similar in all mice, peaked at day 4 postinfection (Fig. 2A) and was comparable, when PBL (Fig. 2A) and spleens (data not shown) were analyzed. However, because T cell expansion not only accompanies activation but also tolerization of T cells (24, 25), we determined diabetes induction as a sign for CTL activation in vivo. Although all mice treated with mature or semimature DC were clinically scored as diabetic (Fig. 2B, left panel), the urine glucose levels in the group treated with TNF-DC was in average 2- to 3-fold lower (Fig. 2B, left panel). In addition, while LPS-DC-treated animals did not clinically improve, the TNF-DC-treated mice fully recovered from diabetes after 27 days (Fig. 2B, left panel). However, after complete recovery of the animals vaccinated with TNF-DC (day 27, Fig. 2B, left panel), a second immunization of the recovered animals with OVA257–264 (100 μg in CFA s.c.) induced strong diabetes with no recovery and, therefore, excluded the possibility that CD8 T cell tolerance induction by TNF-DC could be accompanied by transient acquisition of cytolytic function (Fig. 2C). Taken together, these data suggest that semimature TNF-DC have no direct tolerogenic capacity toward CD8+ T cells, but rather activate them to effector killer cells. However, this activation was weaker as compared with CTL induced by fully mature LPS-DC.
FIGURE 2. Semimature DC induce cytotoxicity in transgenic CD8+ OT-I T cells. A, RIP-OVAlow mice received 1 x 106 CD8+ OT-I T cells i.v. and were immunized 1 day later with 3 x 106 either LPS- or TNF--matured, OVA257–264-loaded DC or not immunized. A, Representative flow cytometry dot plots of blood analysis from day 4 after immunization. Cells within the lymphocyte gate of forward scatter/side scatter analysis (data not shown) were first analyzed for expression of CD8 (data not shown). CD8+ lymphocytes were further analyzed for expression of the transgenic TCR-V segments 2 and 5.1, expressed by OT-I T cells (A, dot plots). The mean percentage of OT-I T cells within CD8+ peripheral blood lymphocytes and the SDs are indicated for every treatment group (n = 6). The kinetics of OT-I T cell expansion after immunization with the different DC is shown in the right graph. Mice from this experiment were also monitored for induction of diabetes by measuring glucosuria. B, Upper left panel, course of diabetes development in Rip-OVAlow mice after immunization with differently matured, OVA257–264-loaded DC. In the lower left panel, the same data are displayed as percent of diabetic mice. Right panel, In this experiment RIP-OVAlow mice received both 1 x 106 (CD8+) OT-I and 2.8 x 106 (CD4+) OT-II cells and DC for immunization were pulsed with both OVA257–264 (MHC I) and OVA323–339 (MHC II). In the upper graph, the course of diabetes development after immunization is shown; the lower graph displays the same data as percent of diabetic mice (n = 4 mice per group). C, Mice from TNF-DC and control groups from the experiment shown in (B, left panel) without signs of diabetes at day 27 after DC-vaccination were immunized with 100 μg of OVA257–264 (MHC I)-peptide emulsified in CFA s.c. to determine whether remaining OT-I cells were tolerant. The course of diabetes development was monitored.
It has previously been shown, that TNF-DC can mediate tolerance via IL-10-producing Ag-specific CD4 regulatory T cells (14). Because RIP-OVAlow mice are tolerant for OVA and their repertoire of T cells specific for OVA is either purged or tolerized (16, 21, 22), we studied whether provision of an exogenous source of OVA-specific CD4 Th cells from TCR-transgenic OT-II mice might overcome the absence of tolerance induction by TNF-DC.
Therefore, we cotransferred both OT-I and OT-II cells into RIP-OVAlow mice and treated them with LPS-DC or TNF-DC pulsed simultaneously with both OVA257–264 and OVA323–339. However, also in the presence of CD4+ OT-II cells and their specific peptide on DC, mice from both experimental groups developed diabetes (Fig. 2B, right panel). Moreover, mice immunized with OVA257–264/OVA323–339-loaded TNF-DC did not recover from diabetes, and their urine glucose levels were similarly elevated as compared with mice immunized with fully mature DC (Fig. 2B, right panel). So far our data indicate that TNF-DC are neither capable without (Fig. 2B) or with (Fig. 2B) the help of Ag-specific CD4 cells to tolerize OVA-specific OT-I CD8 T cells and rather activate CD8 cells to become cytotoxic effector cells.
Semimature DC induce CTL function in C57BL/6 mice and simultaneously protect from CD4 T cell-mediated EAE
To exclude that absence of tolerance induction by TNF-DC was due to the transgenic nature of OT-I or OT-II T cells, or their artificially elevated precursor frequencies, we next compared polyclonal CTL responses induced by mature and semimature DC. To analyze the cytotoxic activity induced by the different DC, we performed "in vivo killer assays" (20) in normal C57BL/6 mice. DC-treated mice were injected with a 1:1 mixture of differently CFSE-labeled spleen cell targets, either pulsed with OVA257–264-peptide (CFSE high) or unpulsed (CFSE low) (Fig. 3A). CTLs generated by immunization of the mice with mature OVA257–264-loaded DC, selectively killed the peptide-loaded (CFSE high) target cell population (Fig. 3A) within a few hours, allowing calculation of specific lysis (20). As shown in Fig. 3A, killing of injected OVA257–264-pulsed spleen cells was not statistically different in mice treated with OVA257–264-loaded LPS- or TNF-DC. Furthermore, when CD4 T cell epitopes of MOG or OVA or both were provided together with the OVA257–264 MHC I epitope on DC as well as on the targets, similar specific killing was observed in each group (Fig. 3A). These results indicate, that neither OVA- nor MOG-specific IL-10-producing CD4 T cells (Fig. 1) suppressed the CD8 T cell responses as bystanders.
We next analyzed cytokine production of CD8 T cells induced by mature or semimature DC. Restimulated splenocytes from mice treated with OVA257–264-pulsed LPS-DC produced mainly IFN-, but no or little IL-10, while semimature TNF-DC did not elicit any significant cytokine responses at all (Fig. 3B). The absence of IFN- secretion by CD8 T cells from mice immunized with TNF-DC (Fig. 3B) in contrast to those treated with fully mature DC indicates incomplete induction of CTL responses in vivo. However, these IFN- negative CD8 T cells do show cytolytic functions (Figs. 2B and 3A).
To proof that TNF-DC loaded with either MOG35–55 or both OVA257–264 and MOG35–55 were similarly tolerogenic for MOG-specific T cells and able to protect mice from EAE, we induced EAE in mice treated as in Fig. 3A. As published before (14), immunization with semimature MOG35–55 -loaded DC protected C57BL/6 mice from EAE, while mice without DC-treatment or injected with mature MOG35–55-pulsed LPS-DC developed severe EAE (Fig. 3C, MOG). However, the same semimature DC loaded with both, MOG35–55 and OVA257–264, that induced specific CTLs against the MHC I peptide OVA257–264 (Fig. 3A) protected mice from MOG-induced EAE (Fig. 3C). These data suggest that the same semimature DC, which induce MOG-specific IL-10-producing CD4 (regulatory) T cells triggers simultaneously cytotoxic effector T cells, which kill targets (Fig. 3A) in absence of IFN- production. The latter were not inhibited by bystander regulatory T cells triggered most likely in their vicinity, because efficient cytolytic activity was induced also by double-, and triple-loaded semimature DC (Fig. 3B).
Taken together, these data suggest that the tolerizing capacity of semimature DC is restricted to Ag-specific CD4 T cells, but ineffective to suppress TCR transgenic (Fig. 2) as well as polyclonal CD8 T cells (Fig. 3) specific for OVA.
Discussion
In vivo counterparts to the in vitro-generated TNF-DC used in this study could be DC in the lymphatics, which migrate from nonlymphoid organs to peripheral lymph nodes, transporting apoptotic cells (26) or tissue (self-)Ag (8, 9, 27). In the lymph node and spleen, these migrated DC are crucial to induce peripheral tolerance (2, 28, 29). The semimature stage of these "steady-state" DC might be attributed to lack of activation by pathogens via TLR ligation leading to phenotypically mature DC that do not produce IL-12 (data not shown and Ref. 14). However, also active suppression of IL-12 production in mature DC has been described for helminth Ag, such as schistosome egg Ag (30) and induction of a general block of DC-maturation was demonstrated for secreted Ag from M. tuberculosis (31). Therefore, several arguments emerged, which support the existence of semimature DC in vivo induced either by lack of appropriate signals or pathogen-induced maturation blockade. It is therefore important to understand the functions of such semimature DC in more detail.
In vitro-cultured TNF--treated semimature DC loaded with MOG35–55 peptide induce MOG35–55-specific tolerance and protect C57BL/6 mice from MOG35–55-induced EAE (Ref. 14 and Fig. 3C). As experiments with semimature DC from IL-10–/– mice showed, DC do not have to produce IL-10 themselves to excert tolerogenic capacity (14). This is in contrast to IL-10-producing DC protecting from EAE (32, 33) or inducing mucosal tolerance (11) in other studies. Similar to immature human DC (34, 35) the mechanism of MOG-tolerance induction by semimature DC was revealed to be dependent on IL-10-producing regulatory T cells (14), a mechanism that has been demonstrated to act also through bystander suppression (36, 37).
Therefore, we investigated the tolerizing capacity of semimature DC to CD4 T cells specific for nonself Ag and CD8+ T cells in addition to the described MOG-self Ag specific CD4 T cell response (14). We show that semimature DC induce a similar IL-10-biased cytokine pattern in foreign Ag (OVA)-specific CD4 T cells as observed for self-Ag-specific MOG T cells (Fig. 1).
When adoptively transferred transgenic OT-I or endogenous polyclonal CD8+ T cells were exposed to MHC I specific peptide-loaded semimature DC in vivo, T cell expansion (Fig. 2) as well as cytotoxic T cell effector functions (Figs. 2 and 3) were induced. Also the addition of exogenous IL-10-producing CD4 T cells (Fig. 2) and simultaneous presentation of corresponding CD4 peptides by semimature DC (Figs. 2 and 3) did not lead to tolerance induction, but rather enhanced CD8 T cell responses (Fig. 2). The magnitude of CD8 responses (Fig. 2) was lower and CD8 cytokine production (Fig. 3) absent, when the responses were initiated by semimature DC. This phenomenon could 1) be due to absent IL-12 production by DC (Ref. 14 and data not shown) and 2) be overcome by addition of exogenous Ag-specific CD4+ T cell help (Fig. 2). Similar to another report (10), where presence of CD4+ T cells altered tolerogenic DC in activating DC, we detected an enhancement of CD8 T cell immunity (Fig. 2). When DC were simultaneously loaded with the tolerogenic MOG35–55 peptide as well as OVA peptides, the same DC induced tolerance for MOG35–55 (Fig. 3C), but immunity for OVA by triggering OVA-specific CD8+ cytotoxic T cell responses (Figs. 2B and 3A). This latter experiment indicated that IL-10-producing MOG- or OVA-peptide specific regulatory CD4 T cells were insufficient to exert bystander suppression on CD8 T cells, neither during priming by the same double- or triple (MOG/OVA)-loaded DC, nor at the level of recognition of (MOG/OVA)-presenting targets (Fig. 3). Data from other experimental systems have suggested that bystander suppression is mediated through IL-10, TGF- or cell-cell contacts (36, 37). Hypothetically the semimature DC induce or expand IL-10-producing MOG-specific regulatory T cells, which then may suppress autoreactive clones of similar specificities at other sites and at later time points (for example, when the disease is induced). However, it is not clear yet whether CD8 T cells have similar requirements for tolerization like MOG-specific EAE-inducing CD4 T cells. Therefore, it is possible that semimature DC cannot tolerize CD8 T cells efficiently, suggesting differential requirements for CD4 and CD8 T cells. Another potential reason for the observed differences between CD4 and CD8 T cells might be the differential loading efficiency of TNF-DC with MHC I vs II peptides. In general, much more peptide is required to trigger the effector function of CD4 T cells when compared with CD8 T cells, and this could also account for the phenotype described in this study. Although we tried in our experiments to work under saturating conditions for peptide loading, an adequate comparison of MHC I- and II-induced responses seems to be impossible.
The OVA-specific CTL induced by TNF-DC were unable to produce IFN-, which is an indication that their activation was not complete (Fig. 3B). Therefore, together with the data from Fig. 2, one could speculate that a reason for this observed cytotoxic activity could be the relatively low threshold signals needed to trigger CTL-activity: one or few specific TCR/peptide-MHC interactions are necessary to induce the cytolytic machinery (38, 39) and even partially activated CD8 T cells can gain cytoloytic activity (40, 41). In earlier studies the phenomenon of cytotoxic activity in absence of cytokine production has been called "split anergy" (42, 43). Taken together, "CTL" induced by IL-12– TNF-DC might be severely impaired in their tumor- (44) and virus-clearing (45) efficiencies, but are definitely not tolerized.
Overall our results suggest that one DC can exert different functions on CD4 and CD8 T cells simultaneously: induction of CD4 T cell tolerance as well as CD8 T cell immunity. However, it remains an open question, if such an artificially enforced experimental situation of a double-peptide-loaded semimature DC is at all existing in vivo. A potential restriction for such a scenario could be the intrinsically different Ag uptake, processing, and presentation capacities of distinct DC-subsets; i.v. injected soluble OVA has been shown to be taken up by splenic DC, which after isolation, markedly differed in their capacity to present the processed OVA. Although the CD8+ DC predominantly cross-presented OVA to CD8 T cells but showed little stimulation of CD4 T cells, the contrary was observed for the CD8– DC subsets which readily presented OVA in an MHC II-restricted fashion but poorly on MHC I molecules (46).
However, other examples for a segregation of function within one APC have been published before. It was shown that DC copulsed with microbial and helminth Ag were able to prime Th1 responses for the microbial Ag as well as Th2 responses for the helminth Ag (47). Therefore, there is increasing evidence for differential handling of different Ag within the same APC. As a consequence, subsequent induction of qualitatively different immune responses by one and the same DC could appear.
Specifically relevant for the question of maintenance of tolerance to apoptotic self-material were recent findings demonstrating that apoptotic (self-) material and bacterial Ag are handled by discrete intracellular phagosomal maturation and processing pathways in APC, depending on their respective TLR signaling (48). Such a signaling by bacterial material induced phagocytosis and processing pathways, which were independently regulated from constitutive pathways for apoptotic material (48). It seems an attractive hypothesis that DC are equipped with endogenous control mechanisms, allowing them to integrate processing and presentation requirements for different Ags to avoid in vivo situations like the one induced by unphysiological external peptide loading in our study.
Although the physiological occurrence of such a simultaneous induction of tolerance and immunity in vivo seems unlikely, our findings have implications for in vitro-loaded DC as they are currently used for human cancer immunotherapy. When patient-derived DC are loaded with peptides on MHC I and/or II molecules but are improperly matured, the possibility of simultaneous CD4 tolerance and only partial CD8 immunity should be taken into consideration. In addition, these data indicate that CD4 and CD8 T cells have different activation requirements from the side of the APC. Althoughthe semimature state of the presenting DC is suboptimal and therefore still tolerogenic for CD4 T cells, the costimulatory and cytokine equipment is sufficient to induce CTL activity, however, with reduced/absent IFN- production.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank A. Bol and W. Mertl for expert animal facility maintenance.
Footnotes
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.
1 This work was supported by the Deutsche Forschungsgemeinschaft Grant Br 1889/2, SFB571, SFB456 and Wilhelm-Sander Stiftung for T.B.; and Grant LU 851/2-1 and the SFB643 (to M.B.L.); and the ELAN program of the University Hospital Erlangen (to C.W.).
2 P.K. and C.W. contributed equally.
3 M.B.L. and T.B. contributed equally.
4 Address correspondence and reprint requests to Dr. Thomas Brocker, Institute for Immunology, Ludwig-Maximilians-Universit?t München, Goethestrasse 31, 80336 München, Germany. E-mail address: tbrocker@ifi.med.uni-muenchen.de
5 Abbreviations used in this paper: DC, dendritic cell; MOG, myelin oligodendrocyte glycoprotein; MHC II, MHC class II; MHC I, MHC class I; EAE, experimental autoimmune encephalomyelitis.
Received for publication September 7, 2004. Accepted for publication December 6, 2004.
References
Banchereau, J., R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392:245.
Hawiger, D., K. Inaba, Y. Dorsett, M. Guo, K. Mahnke, M. Rivera, J. V. Ravetch, R. M. Steinman, M. C. Nussenzweig. 2001. Dendritic cells induce peripheral T cell unresponsiveness under steady state conditions in vivo. J. Exp. Med. 194:769.
Bonifaz, L., D. Bonnyay, K. Mahnke, M. Rivera, M. C. Nussenzweig, R. M. Steinman. 2002. Efficient targeting of protein antigen to the dendritic cell receptor DEC-205 in the steady state leads to antigen presentation on major histocompatibility complex class I products and peripheral CD8+ T cell tolerance. J. Exp. Med. 196:1627.
Probst, H. C., J. Lagnel, G. Kollias, M. van den Broek. 2003. Inducible transgenic mice reveal resting dendritic cells as potent inducers of CD8+ T cell tolerance. Immunity 18:713.
Wilson, N. S., D. El-Sukkari, G. T. Belz, C. M. Smith, R. J. Steptoe, W. R. Heath, K. Shortman, J. A. Villadangos. 2003. Most lymphoid organ dendritic cell types are phenotypically and functionally immature. Blood 102:2187.
De Smedt, T., B. Pajak, E. Muraille, L. Lespagnard, E. Heinen, P. De Baetselier, J. Urbain, O. Leo, M. Moser. 1996. Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J. Exp. Med. 184:1413.
Steinman, R. M., M. C. Nussenzweig. 2002. Avoiding horror autotoxicus: the importance of dendritic cells in peripheral T cell tolerance. Proc. Natl. Acad. Sci. USA 99:351.
Scheinecker, C., R. McHugh, E. M. Shevach, R. N. Germain. 2002. Constitutive presentation of a natural tissue autoantigen exclusively by dendritic cells in the draining lymph node. J. Exp. Med. 196:1079.
Turley, S., L. Poirot, M. Hattori, C. Benoist, D. Mathis. 2003. Physiological cell death triggers priming of self-reactive T cells by dendritic cells in a type-1 diabetes model. J. Exp. Med. 198:1527.
Albert, M. L., M. Jegathesan, R. B. Darnell. 2001. Dendritic cell maturation is required for the cross-tolerization of CD8+ T cells. Nat. Immunol. 2:1010
Akbari, O., R. H. DeKruyff, D. T. Umetsu. 2001. Pulmonary dendritic cells producing IL-10 mediate tolerance induced by respiratory exposure to antigen. Nat. Immunol. 2:725.
Salomon, B., D. J. Lenschow, L. Rhee, N. Ashourian, B. Singh, A. Sharpe, J. A. Bluestone. 2000. B7/CD28 costimulation is essential for the homeostasis of the CD4+CD25+ immunoregulatory T cells that control autoimmune diabetes. Immunity 12:431
Yamazaki, S., T. Iyoda, K. Tarbell, K. Olson, K. Velinzon, K. Inaba, R. M. Steinman. 2003. Direct expansion of functional CD25+ CD4+ regulatory T cells by antigen-processing dendritic cells. J. Exp. Med. 198:235.
Menges, M., S. Rossner, C. Voigtlander, H. Schindler, N. A. Kukutsch, C. Bogdan, K. Erb, G. Schuler, M. B. Lutz. 2002. Repetitive injections of dendritic cells matured with tumor necrosis factor induce antigen-specific protection of mice from autoimmunity. J. Exp. Med. 195:15.
Lutz, M. B., G. Schuler. 2002. Immature, semi-mature and fully mature dendritic cells: which signals induce tolerance or immunity?. Trends Immunol. 23:445.
Blanas, E., F. R. Carbone, J. Allison, J. F. Miller, W. R. Heath. 1996. Induction of autoimmune diabetes by oral administration of autoantigen. Science 274:1707.
Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, F. R. Carbone. 1994. T cell receptor antagonist peptides induce positive selection. Cell 76:17.
Barnden, M. J., J. Allison, W. R. Heath, F. R. Carbone. 1998. Defective TCR expression in transgenic mice constructed using cDNA-based - and -chain genes under the control of heterologous regulatory elements. Immunol. Cell Biol. 76:34.
Lutz, M. B., N. Kukutsch, A. L. Ogilvie, S. Rossner, F. Koch, N. Romani, G. Schuler. 1999. An advanced culture method for generating large quantities of highly pure dendritic cells from mouse bone marrow. J. Immunol. Methods 223:77.
Coles, R. M., S. N. Mueller, W. R. Heath, F. R. Carbone, A. G. Brooks. 2002. Progression of armed CTL from draining lymph node to spleen shortly after localized infection with herpes simplex virus 1. J. Immunol. 168:834.
Kurts, C., R. M. Sutherland, G. Davey, M. Li, A. M. Lew, E. Blanas, F. R. Carbone, J. F. Miller, W. R. Heath. 1999. CD8 T cell ignorance or tolerance to islet antigens depends on antigen dose. Proc. Natl. Acad. Sci. USA 96:12703.
Kurts, C., J. F. Miller, R. M. Subramaniam, F. R. Carbone, W. R. Heath. 1998. Major histocompatibility complex class I-restricted cross-presentation is biased towards high dose antigens and those released during cellular destruction. J. Exp. Med. 188:409.(Petra Kleindienst, Carste)