Ms. Van Duivenvoorde and Dr. Louis-Plence contributed equally to this work.
Antigen-specific immunomodulation of collagen-induced arthritis with tumor necrosis factor–stimulated dendritic cells
Article first published online: 8 OCT 2004
Copyright © 2004 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 50, Issue 10, pages 3354–3364, October 2004
How to Cite
van Duivenvoorde, L. M., Louis-Plence, P., Apparailly, F., van der Voort, E. I. H., Huizinga, T. W. J., Jorgensen, C. and Toes, R. E. M. (2004), Antigen-specific immunomodulation of collagen-induced arthritis with tumor necrosis factor–stimulated dendritic cells. Arthritis & Rheumatism, 50: 3354–3364. doi: 10.1002/art.20513
- Issue published online: 8 OCT 2004
- Article first published online: 8 OCT 2004
- Manuscript Accepted: 4 JUN 2004
- Manuscript Received: 23 DEC 2003
- Dutch Arthritis Foundation. Grant Numbers: RF 01-2-401, RF 02-1-402
Dendritic cells (DCs) are crucial for the initiation of T cell immunity and therefore play an important role in the initiation and regulation of immune responses in arthritis. Full mobilization of effector T cells depends on the proper maturation of DCs. Current evidence indicates that the type of T cell response induced is crucially dependent on the activation status of the DCs. In this study, we explored the immunologic effects of differentially matured DCs on the development of collagen-induced arthritis (CIA).
Bone marrow–derived DCs were cultured in the presence of granulocyte–macrophage colony-stimulating factor (GM-CSF). Before immunization with bovine type II collagen (CII) protein, mice were repeatedly injected with DCs that had been pulsed with CII. Immature, semimature, or fully mature DCs were injected. Mice were boosted on day 21 after CII immunization, and the disease course was monitored.
While vaccination with immature or lipopolysaccharide-activated DCs had no significant effect on the disease course, administration of antigen-loaded, tumor necrosis factor (TNF)–modulated DCs propagated in GM-CSF with or without interleukin-4 resulted in a delayed onset of arthritis and a lower clinical score. The response was antigen-specific, since TNF-treated DCs pulsed with a control antigen did not modify the disease course. A specific decrease in the collagen-specific “Th1-associated” IgG2a response was observed, whereas IgG1 titers were unaffected.
CIA can be prevented through vaccination with TNF-matured DCs in an antigen-specific manner. These findings provide a rationale for immunotherapy using DCs in rheumatoid arthritis.
Rheumatoid arthritis (RA) is an autoimmune disease characterized by chronic inflammation and synovial infiltration of immune cells that display an activated phenotype, such as activated CD4+ T cells, B cells, and antigen-presenting cells (e.g., dendritic cells [DCs] and macrophages). While the symptoms associated with RA are well described, the factors that are crucially involved in the induction and/or progression of RA are poorly defined. Nonetheless, it is generally accepted that cells belonging to the adaptive immune system (i.e., B cells and T cells) are intimately engaged in the processes responsible for the induction and/or progression of RA. Since these cells require triggering of their antigen receptors before they can exert their functions, it is likely that antigen recognition is important during the initiation and/or perpetuation of disease.
Unfortunately, no antigens have yet been identified that are involved in the processes underlying the pathology observed in RA patients. It is therefore not presently feasible to use specific antigens to treat RA patients in a disease-specific manner. Likewise, it is currently not possible to prevent in an antigen-specific manner the development of RA in individuals at risk, such as those with undifferentiated polyarthritis. A disease-specific intervention is attractive because such a treatment modality would most likely not induce the side effects that are associated with many of the treatments currently available in clinical practice. Nonetheless, it is not unreasonable to anticipate that such antigen-specific interventions will be an option in the future as a consequence of the rapidly and exponentially increasing knowledge and understanding of the molecular and immunologic principles that underlie RA.
To gain a better appreciation of the feasibility of a disease-specific intervention in RA, we studied an experimental animal model of RA, collagen-induced arthritis (CIA). CIA can be induced in genetically susceptible mouse strains by immunization with bovine type II collagen (CII) protein in Freund's complete adjuvant (CFA). Although this mouse model is certainly not similar to RA, many important features, such as the systemic nature of CIA, the production of autoantibodies, the chronic and destructive inflammation of joints, as well as the associations with the major histocompatibility complex (MHC) highly resemble features of RA (1). Consequently, it is presumed that the most prominent immunologic and inflammatory mechanisms that are operative in RA are also present in CIA. We therefore studied the possibility of modulating the immune response, which leads to chronic inflammatory polyarthritis after injection of CII in CFA, by vaccination with well-defined populations of DCs, since DCs are directly involved in the activation/tolerization of the immune system (2, 3).
DCs are professional antigen-presenting cells that are present in low numbers in all body tissues (4). These cells are specialized in the uptake, transport, processing, and presentation of (self) antigens to T cells. DCs can be divided into 2 populations, immature and mature. Immature DCs are capable of antigen uptake. After an activation stimulus from the innate immune system through, for example, Toll-like receptor triggering, or from the adaptive immune system via CD40 signaling, DCs become mature, as evidenced by an up-regulation of MHC molecules and costimulatory molecules, such as CD40, CD80, and CD86. These mature DCs are no longer capable of antigen uptake but are endowed with the capacity to initiate antigen-specific T cell responses. In contrast, immature DCs are thought to induce antigen-specific tolerance via the induction of regulatory T cells or the deletion of antigen-specific T cells. Thus, DCs play a pivotal role in orchestrating the immune response, since the activation status of DCs imposes an important regulatory control in the induction and tolerization of immune responses against self and non-self antigens.
The realization that DCs may be responsible for T cell polarization and tolerance induction has initiated several studies aimed at inhibiting the unwanted immune responses responsible for organ rejection or graft-versus-host disease and autoimmunity (5–12). Nonetheless, the maturation status of DCs, whether immature, “semimature,” or “alternatively” matured, that is required for optimal tolerance induction in the presence of immunosuppressive agents such as interleukin-10 (IL-10) or dexamethasone is not well defined. It is conceivable that the dynamics and features of different autoimmune diseases in combination with the characteristics of different DC populations determine the outcome of intervention.
In the present study, we investigated the capacity of immature DCs, tumor necrosis factor (TNF)–modulated DCs, and fully mature DCs to initiate tolerance in mice against CIA. TNF-modulated DCs have previously been shown to prevent experimental autoimmune encephalomyelitis (EAE) in C57BL/6 mice (13). Unlike CIA, EAE is an organ-specific autoimmune model in which B cell immunity does not play a crucial role (14). We found that TNF-modulated DCs, but not immature or lipopolysaccharide-matured DCs, can prevent CIA when loaded with CII antigen. The protection was antigen-specific, since only TNF-modulated DCs loaded with CII (protein or peptide) could prevent CIA. Moreover, vaccination with TNF-modulated DCs was associated with an alteration of the IgG2a:IgG1 ratio of CII-specific antibody titers. These data show that targeting DCs can be a powerful way to modulate arthritis in an antigen-specific manner.
MATERIALS AND METHODS
DBA/1J mice were obtained from our own breeding colonies. TBC mice (bovine CII–specific, T cell receptor–transgenic mice) were kindly provided by Dr. W. C. Ladiges (Department of Comparative Medicine, School of Medicine, University of Washington, Seattle, WA) (15, 16). Mice used in the studies were 7–10 weeks old and were maintained in accordance with the national guidelines for animal care.
Induction of CIA and evaluation of arthritis.
CIA was induced in male DBA/1J mice (H-2q). Bovine CII protein (Chondrex, Redmond, WA, or Sigma-Aldrich, l'Isle d'Abeau, France) was dissolved in 0.1M acetic acid solution overnight at 4°C at a concentration of 2 mg/ml. The dissolved CII (100 μg of CII/mouse) was emulsified with an equal volume of CFA (Difco, Detroit, MI, or Perbio Science, Bezons, France), and 100 μl was injected subcutaneously into the base of the tail. This immunization was boosted 3 weeks later with a subcutaneous injection of 100 μg of CII emulsified in Freund's incomplete adjuvant (Difco or Perbio Science).
Beginning 3 weeks after immunization, mice were examined 3 times each week for signs of arthritis. The presence of arthritis in the paws was determined by macroscopic examination. Arthritis severity in each paw was graded according to an established scoring system: 0 = normal joints, 1 = 1 or 2 swollen joints, 2 = >2 swollen joints, and 3 = extreme swelling of the entire paw and/or ankylosis. An arthritis score for each mouse was calculated by summing the scores for each paw. When a mouse had 2 paws with a maximum score of 3, it was euthanized because of ethical considerations, as defined by the local ethics committee. The last score measured at the time of euthanization was the arthritis score for that mouse.
Hind paws were collected for histologic examination, fixed in 4% paraformaldehyde, decalcified in acid-based DC3 (Labonord, Templemars, France), embedded in paraffin, and 5-μm sections were cut. Sections were stained with hematoxylin and eosin–Safranin O. Paw sections were examined by 2 independent observers (PL-P and FA) who were blinded to the experimental group. Four successive tissue sections of 2 different areas of each paw (joints of the midfoot and hindfoot) were scored for overall histopathologic features: 0 = normal, 1 = inflammatory infiltrates and synovial hyperplasia, 2 = pannus formation and cartilage erosion, and 3 = important cartilage erosion and bone destruction.
Preparation of bone marrow–derived DCs.
DCs were generated from bone marrow obtained from DBA/1J mice according to 2 previously described procedures (17, 18). Briefly, bone marrow cells were cultured for 10 days in medium containing granulocyte–macrophage colony-stimulating factor (GM-CSF) (17, 19). After 10 days of culture, DCs were removed and used as immature DCs, or they were activated with 500 units/ml of TNF (Tebu-Bio, Heerhugowaard, The Netherlands) or with 1 μg/ml of lipopolysaccharide (LPS; Sigma-Aldrich, Zwijndrecht, The Netherlands). In some experiments, as indicated below, DCs were propagated from bone marrow progenitor cells in medium containing both recombinant mouse GM-CSF (2.5 ng/ml) and recombinant mouse IL-4 (5 ng/ml) (PeproTech, Le Perray-en-Yvelines, France) for 7 days. Maturation of DCs was induced on day 6 by transferring DCs into new 6-well plates with 500 units/ml of TNF (R&D Systems, Abingdon, UK) for 24 hours.
Antibodies and FACS analysis.
The following antibodies were purchased from BD PharMingen (Alphen aan den Rijn, The Netherlands): fluorescein isothiocyanate (FITC)–labeled anti-CD11c, FITC-labeled anti-CD11b, phycoerythrin (PE)–labeled anti-CD40, PE-labeled anti-CD80, FITC-labeled anti-CD86, PE-conjugated anti–class II (clone M5/115), and FITC-labeled isotype controls. Immature DCs and DCs activated for 24 or 48 hours with 500 units/ml of TNF or 1 μg/ml of LPS were incubated with the appropriate antibodies for 15 minutes at 4°C in the dark. After 3 washing steps, the cells were fixed in 0.5% paraformaldehyde. Flow cytometry was performed with a FACSCalibur machine (Becton Dickinson, Mountain View, CA).
We used nonactivated DCs or DCs that had been activated overnight with 500 units/ml of TNF or 1 μg/ml of LPS. Supernatants were tested for IL-12p40 content using a standard sandwich enzyme-linked immunosorbent assay (ELISA). The coating antibody was rat anti-mouse IL-12p40/p70 monoclonal antibody (clone C15.6; BD PharMingen). The detection antibody was biotinylated rat anti-mouse IL-12p40/p70 (clone C17.8; BD PharMingen). Streptavidin–horseradish peroxidase (HRP) (Sanquin, Amsterdam, The Netherlands) and ABTS (Sigma-Aldrich, Zwijndrecht, The Netherlands) were used as enzyme and substrate, respectively.
Induction of collagen-specific T cell responses in vitro.
Spleen and lymph nodes were isolated from a TBC mouse. T cells were purified using nylon wool and anti-B220 antibodies (hybridoma 6B2; American Type Culture Collection, Rockville, MD) with sheep anti-rat beads (Dynal Biotech, Hamburg, Germany). T cells were stimulated with antigen-loaded DCs (immature, TNF-activated, or LPS-activated) and cultured in Iscove's modified Dulbecco's medium (Cambrex Bioscience, Verviers, Belgium) containing 8% heat-inactivated fetal calf serum (Bodinco, Alkmaar, The Netherlands), 100 units/ml of penicillin, 2 mML-glutamine, and 20 μM β-mercaptoethanol in 96-well round-bottomed plates at a concentration of 100,000 T cells/well. Proliferation was measured 3 days later, after the addition of 0.5 μCi/well of 3H-thymidine. Values reported represent the average of triplicate cultures with the values for the medium subtracted.
Stimulation of CD4+ T cells for cytokine production.
Naive, nylon-wool–purified DBA/1J T cells (H-2q) were constitutively stimulated with allogeneic DCs (D1 cell line [H-2b]; kindly provided by Dr. P. Ricciardi-Castagnoli, Department of Biotechnology and Bioscience, University of Milan–Bicocca, Milan, Italy) as described previously (20). Briefly, T cells were cultured with either 500 units/ml of TNF or 1 μg/ml of LPS-activated D1 cells and restimulated on day 6 after stimulation (total of 3 times). Cells were then harvested, counted, and stimulated with 2 μg/ml of coated anti-CD3 (clone 145-2C11) for 48 hours at 37°C.
Supernatants were tested for interferon-γ (IFNγ) and IL-5 content using a standard sandwich ELISA. Coating antibodies were rat anti-mouse IFNγ (clone R4-6A2) and rat anti-mouse IL-5 (clone TRFK5). Detection antibodies were biotinylated rat anti-mouse IFNγ (clone XMG1.2) and biotinylated rat anti-mouse IL-5 (clone TRFK4). All antibodies were purchased from BD PharMingen. Streptavidin–HRP and ABTS were used as enzyme and substrate, respectively.
Treatment of mice with DCs.
For in vivo experiments, DCs were stimulated for 4 hours with either 500 units/ml of TNF or 1 μg/ml of LPS and loaded with 10 μg/ml of bovine CII peptide (amino acids 256–270) or DCs were pulsed with CII protein (2 μg/ml), and treated with TNF for 24 hours as indicated below. Mice were given 3 intravenous injections (on days –7, –5, and –3 prior to immunization) of 2.5 × 106 DCs in 200 μl of phosphate buffered saline (PBS).
Measurement of antigen-specific antibodies in serum.
Antibodies were measured using a standard sandwich ELISA. Immuno-Maxisorp 96-well plates (Nunc, Roskilde, Denmark) were coated overnight at 4°C with 2 μg/ml of murine CII protein (Chondrex) or 5 μg/ml of purified protein derivative (PPD; a component of CFA) (Statens Seruminstitut, Copenhagen, Denmark). After washing with PBS–0.5% Tween 20, plates were blocked with PBS–10% milk for 2 hours at 4°C. Serially diluted mouse serum was then incubated overnight at 4°C on the washed plates. Plates were subsequently treated with 1 of the following detection antibodies: HRP-conjugated anti-mouse IgG1, or HRP-conjugated anti-mouse IgG2a (Southern Biotechnology, Birmingham, AL). Detection was performed using 3,3′,5,5′-tetramethylbenzidine as substrate (Sigma-Aldrich, Zwijndrecht, The Netherlands).
Antibody units for both murine CII and PPD were determined using a reference serum created from pooled sera of arthritic mice. This serum pool was assigned an arbitrary level of 100 units of antigen-specific antibodies.
Differences in disease severity and in antibody production were analyzed with the nonparametric Mann-Whitney U test or Student's t-test, as appropriate, according to the distribution of the data. P values less than 0.05 with a 95% confidence interval were considered significant. Percentages were compared using the chi-square test.
Influence of TNF on the maturation of DCs.
DCs are thought to play a critical role in setting the immune response against (self) antigens, mediating either immunity or tolerance. The capacity of DCs to modulate immunity in an antigen-specific manner offers new intervention possibilities for preventing or treating autoimmunity without the nonspecific side effects that are associated with most current treatment modalities. To study the potential of DCs to prevent arthritis, we wished to focus on TNF-modulated DCs, since these DCs have been shown to prevent EAE, an organ-specific T cell–mediated autoimmune disease, in C57BL/6 mice (13).
DCs were generated from bone marrow cells. After 10 days of culture in GM-CSF containing medium, the cells were analyzed for surface molecule expression by fluorescence-activated cell sorting (FACS). Figure 1A shows that all cells expressed CD11c and CD11b, indicating that these DCs belong to the myeloid lineage.
To study the response of DCs from DBA/1J mice to various maturation stimuli, DCs were treated with TNF or LPS. Untreated DCs were used as immature controls. After 24 and 48 hours, the cells were harvested and analyzed by FACS for the expression of MHC class II and several costimulatory molecules (Figure 1B). Compared with immature DCs, TNF-treated DCs displayed elevated levels of MHC class II, CD40, CD80, and CD86. Nevertheless, these cells were less mature compared with DCs activated by LPS, which showed more pronounced levels of MHC class II and costimulatory molecules as analyzed after 24 hours. TNF was, however, capable of providing a potent activation stimulus to the DCs, since the expression levels of these molecules were further increased after 48 hours. Likewise, LPS-activated DCs produced high levels of IL-12p40 (Figure 1C), while TNF-treated DCs produced moderate levels. No production of IL-12p40 by immature DCs was measured. Together, these findings indicate that TNF treatment of DBA/1J mouse DCs led to an intermediate, but not fully activated, phenotype since several activation markers were only moderately up-regulated.
Induction of poor T cell proliferation by immature and TNF-matured DCs.
To investigate the T cell–stimulatory capacities of these DC types, immature DCs were matured with either LPS or TNF and incubated with CII protein. After 24 hours, the DCs were cultured with purified collagen-specific CD4+ T cells from TBC mice. After 3 days, T cell proliferation was measured by the incorporation of 3H-thymidine (Figure 2). LPS-activated DCs displayed the highest T cell–stimulating capacity, TNF-modulated DCs displayed a moderate capacity, and immature DCs displayed a low capacity to stimulate CII-specific T cells. Similar findings were obtained when DCs were used as stimulator cells to activate alloreactive T cells (data not shown). Together, these data indicate that TNF treatment of DCs led to a partially activated phenotype, as evidenced by the expression of costimulatory molecules and cytokines and their ability to stimulate T cell proliferation.
Protection against CIA by vaccination with TNF-modulated DCs.
The data described above indicate that TNF-treated DCs are partially activated compared with immature (nonactivated) and LPS-treated (fully activated) DCs. To analyze the capacity of these DC subsets to modulate arthritis induction, immature, TNF-treated, and LPS-treated DCs were pulsed with CII peptide 256–270. These DCs were then injected intravenously at 3 different time points (days –7, –5, and –3) before CIA induction.
The incidence and onset of disease in the groups treated with CII-loaded immature or LPS-matured DCs were similar to those in the control group mice immunized with CII/CFA only. In contrast, mice treated with CII-pulsed, TNF-modulated DCs displayed a lower incidence and a delayed induction of disease as compared with the 3 other groups. The day of disease onset in mice treated with the CII-pulsed, TNF-modulated DCs was 49 ± 7 (mean ± SEM), compared with 34 ± 2 in the control mice (P < 0.05). Similarly, mice vaccinated with TNF-modulated DCs had significantly less-severe disease compared with control mice and mice treated with immature or LPS-activated DCs (Figure 3A). Moreover, 38% (5 of 13) of the mice treated with CII-loaded, TNF-modulated DCs were even protected from CIA for >100 days after CII/CFA immunization, whereas all mice in the other groups developed CIA on day 65 (data not shown).
To rule out the possibility that the observed protective effect was due to the specific propagation protocol used for preparing the DCs, we also generated for comparison TNF-matured DCs that were expanded in the presence of recombinant GM-CSF plus recombinant IL-4. Again, TNF maturation conferred upon DCs the capacity to modulate disease outcome, as evidenced by a significant delay in disease onset: day 42 ± 5 and day 40 ± 8 for TNF-treated DCs cultured with GM-CSF or with recombinant GM-CSF plus recombinant IL-4 respectively, compared with day 32 ± 9 for untreated mice. Likewise, compared with controls, the severity was significantly attenuated for the mice treated with TNF-modulated DCs in GM-CSF as well as for the mice injected with TNF-modulated DCs in recombinant GM-CSF plus recombinant IL-4 (Figure 3B). In the control mice, the mean (±SEM) clinical score reached at the end of the experiment was 8.0 ± 1.1, as compared with 3.8 ± 2.3 for mice injected with TNF-stimulated DCs in GM-CSF (P = 0.001) and 4.2 ± 1.4 for mice vaccinated with TNF-activated DCs in recombinant GM-CSF plus recombinant IL-4 (P < 0.0001).
Histologic examination of arthritic joints from the control mice revealed a significantly increased inflammatory synovitis, pannus formation, and massive inflammatory cell infiltration compared with the 2 groups vaccinated with TNF-modulated DCs (Figure 3C–E). This was also reflected by the number of paws with severe histologic scores, since these were significantly lower in TNF-treated DCs propagated in GM-CSF or in recombinant GM-CSF plus recombinant IL-4 (35.7% and 28.6%, respectively) compared with the control group (64.3%) (P < 0.0001). Together, these data indicate that vaccination with TNF-modulated DCs pulsed with CII can protect against CIA.
Immunoregulation of CIA in an antigen-specific manner.
The above data show a significant decrease in arthritis severity and a delay in onset of CIA by CII-loaded, TNF-modulated DCs. To determine whether this effect was antigen-specific, we generated TNF-stimulated DCs and pulsed them with either CII peptide, keyhole limpet hemocyanin protein (as irrelevant antigen), or left them unpulsed. Figure 4 shows that neither unpulsed DCs or KLH-loaded DCs induced the same effect as TNF-activated DCs that had been pulsed with CII peptide. These results indicate that the beneficial effect of CII-pulsed, TNF-modulated DCs is antigen-specific, since it is crucially dependent on presentation by the DCs of the CII antigen.
Decreased titers of CII-specific, Th1-associated antibody after vaccination with TNF-modulated DCs.
The data presented above indicate that the protective effects induced by TNF-modulated DC vaccination is T cell–mediated, since presentation of the CII peptide by the DCs was required. It has previously been shown that B cell–deficient mice are resistant to CIA (21), indicating that collagen-specific antibodies are crucial for disease induction. Since T cells are involved in the generation of optimal B cell immunity, we investigated whether CII-specific B cell responses are affected by vaccination with TNF-activated DCs. Serum was collected at different time points after disease induction, and antibody titers were measured by ELISA.
As shown in Figure 5A, titers of CII-specific IgG1 antibody were similar in all groups. Likewise, there was no significant difference in IgG2a antibody titers in mice treated with immature or LPS-activated DCs as compared with untreated mice. In contrast, when mice were vaccinated with CII peptide–loaded, TNF-modulated DCs, lower titers of CII-specific antibody of the IgG2a isotype were detected (Figure 5A). This resulted also in a markedly reduced IgG2a:IgG1 ratio (Figure 5B). The decrease in IgG2a antibody titers was antigen-specific, since the IgG2a:IgG1 ratio against PPD was unaffected (Figure 5B). Together, these data indicate that treatment with antigen-pulsed, TNF-stimulated DCs modulates CII-specific immunity, resulting in protection against CIA.
Altered cytokine production by CD4+ T cells after stimulation with TNF-activated DCs.
The above data indicate a specific decrease in the titers of IgG2a antibody against CII. Since switching toward the IgG2a isotype is highly associated with a typical Th1 response in mice, we analyzed whether the production of the Th1-associated cytokine IFNγ was altered in mice treated with TNF-modulated DCs. Indeed, spleen cells from mice that had received TNF-modulated DCs displayed a considerably lower ability to produce IFNγ after stimulation (mean ± SEM 1.13 ± 0.31 ng/ml versus 13.28 ± 0.52 ng/ml when mice are only immunized).
To directly address the question of whether TNF-treated DCs are endowed with the capacity to influence the cytokine production profile of T cells, we investigated whether repetitive stimulation by TNF-modulated DCs was associated with a decrease in IFNγ production by T cells. After 3 weeks of culture, polarized T cells were harvested, counted, and activated with anti-CD3. In this way, all T cells were activated in a similar manner, ruling out the possibility that differences in cytokine production are related to a reduced number of antigen-specific T cells or a reduced antigenic stimulus.
Figure 6A shows that IFNγ production, as measured by ELISA, was significantly lower in cultures stimulated with TNF-modulated DCs than in cultures stimulated with LPS-matured DCs. Moreover, T cells cultured with TNF-stimulated DCs showed an enhanced production of IL-5 compared with cultures activated with LPS-matured DCs (Figure 6B). These data further strengthen the concept that repetitive stimulation with TNF-treated DCs does not support the production of IFNγ by T cells, but rather, promotes Th2-associated cytokine production.
In this study, we demonstrated that DCs can profoundly affect the outcome of CIA by altering the underlying CII-directed immune response. Vaccination of mice with TNF-modulated DCs, but not fully activated DCs, led to a selective reduction of CII-specific antibodies of the IgG2a isotype, but not the IgG1 isotype. This effect was antigen-specific, since no alterations in the antibody response against a control antigen were observed. The shift in balance of the IgG2a:IgG1 ratio was correlated with protection against CIA, as reflected in a lower incidence, a later onset, and a lower severity of disease.
These findings can be explained by the ability of IgG2a and IgG1 antibodies to recruit effector mechanisms, such as complement activation and triggering of IgG Fc receptors (FcγR). In a model of autoantibody-mediated anemia using IgG isotype–switch variants of an anti-erythrocyte autoantibody (22), it was found that IgG2a was superior to IgG1 in activating complement. Moreover, it was found that the IgG2a isotype was able to interact very efficiently with FcγR. Accordingly, 20 times higher doses of IgG1 than of IgG2a autoantibodies were required to induce autoantibody-mediated pathology. Both complement and FcγR-mediated mechanisms play a pivotal role in the induction of CIA. Mice deficient in C3 or C5 are resistant to CIA (23, 24). Likewise, mice lacking the FcγR develop less severe arthritis compared with their wild-type littermates (25, 26). Therefore, it is likely that the specific reduction in CII-directed IgG2a antibodies in mice treated with TNF-modulated DCs accounts for the protection against CIA.
These data are important because they show that TNF-activated DCs can also be used to protect mice from autoimmune diseases mediated by autoantibodies. It has previously been shown that vaccination with TNF-modulated DCs can protect against EAE when loaded with the appropriate autoantigenic peptide (13). In contrast to CIA, EAE is an organ-specific autoimmune disease that affects the central nervous system, but it is not crucially dependent on B cells or autoantibodies (27, 28). In the EAE model, it was found that vaccination with peptide-pulsed TNF-stimulated DCs led to the induction of IL-10–producing CD4+ T cells and an inhibited generation of autoantigen-specific Th1 cells. It is therefore anticipated that a similar skewing of the T cell response is at the basis of the effects observed in the CIA model.
The differentiation of naive CD4+ T cells into Th1 or Th2 effector cells is a critical process in the adaptive immune response. Th1 cells produce IFNγ, promote cellular immunity, and are critical for the generation of IgG2a antibodies in mice (29). We showed that repetitive stimulation of T cells by TNF-modulated DCs decreased IFNγ secretion and enhanced IL-5 production, indicating that TNF-stimulated DCs polarize T cells toward a Th2-cytokine production profile. In addition, a decrease in IFNγ secretion was noted in the CIA model when mice were pretreated with CII-loaded, TNF-modulated DCs as compared with mice that were only immunized.
Many factors influence the differentiation process of CD4+ T cells into Th1 or Th2 effector cells, including the dose of the antigen, the strength of the signal through the T cell receptor, and costimulation. However, cytokines have emerged as key determinants of the outcome of this differentiation. IL-12 is critically involved in the promotion of Th1 development by means of a signaling pathway that involves activation of STAT-4. We observed a reduced production of IL-12 by TNF-stimulated DCs compared with LPS-activated DCs, a parameter that likely determines the outcome of immunomodulation by a decreased ability to support the development of Th1. Nonetheless, we did not observe any effect on the outcome of CIA when antigen-loaded, immature DCs were used for vaccination. In addition, these DCs do not produce IL-12, which is evidence that other factors generated after TNF modulation are critical for the effects induced by injection of TNF-stimulated DCs.
TNF-treated DCs display a semiactivated phenotype, as determined by the expression of surface markers. Whether the semiactive status induced by TNF maturation is responsible for the effects observed, or whether specific, but as-yet-unidentified, signals are playing a dominant role is currently not known. With respect to the latter possibility, it is tempting to speculate that specific molecular interactions between DCs and T cells are involved in setting the outcome of T cell responsiveness. For example, it has recently been described that triggering of T cells by T cell immunoglobulin domain, mucin domain–containing molecule 3 (TIM-3), leads to the inhibition of Th1, but not Th2, responses (30). By up-regulation of molecules such as TIM-3 ligand (that has not yet been identified) under the influence of TNF signaling, DCs would be harnessed with the capacity to tune the outcome of T cell reactivity by the inhibition of Th1-responses. Thus, the “plastic” properties of DCs could be explained by the differential regulation of a combination of several ligand/receptor pairs under the influence of different maturation stimuli, allowing the inhibition of various differentiation routes in T cells.
Modulating the immune system in an antigen-specific manner to prevent or treat arthritis might be a very advantageous intervention, since it is likely to induce fewer side effects. Therefore, different treatment strategies aimed at redirecting immune reactivity are currently being explored. A clear disadvantage of using DCs for immunomodulation is the necessity to culture DC populations on an individual basis, thereby preventing the large-scale use of DC-based intervention strategies. Nonetheless, this “ex vivo” approach appears to be valuable because DCs can be loaded with a large range of antigens and, most importantly, DC maturation can be regulated. This could, for example, bypass the potential hazard associated with nasal tolerance induction in cases in which the “tolerogen” is given around the time of an upper respiratory tract infection. This situation could result in DC activation and, as a consequence, immunity rather than tolerance (31). Indeed, several studies using DCs to induce or boost antitumor responses in cancer patients have already been conducted. The tolerizing effects of DCs have also been studied in humans. In one of them, “immature” DCs pulsed with an influenza matrix peptide were injected into healthy volunteers (32). As a result, influenza-specific IFNγ-producing CD8+ T cells initially disappeared from the blood, and in their place, peptide-specific IL-10–secreting T cells emerged. In contrast, injection of fully activated DCs, loaded with the same peptide expanded the influenza-specific immunity. These findings indicate that DCs in different activation states can be used safely in humans to modulate the outcome of immune responses.
Whether the use of antigen-specific immune intervention protocols will be feasible for the successful treatment or modulation of RA remains to be established. No antigens that are casually related to the induction and/or maintenance of RA are currently known. However, it has recently been shown that antibodies against cyclic citrullinated peptides (anti-CCP) predate the onset of RA by several years (33). Moreover, we have shown that the presence of anti-CCP antibodies in patients diagnosed as having undifferentiated polyarthritis is highly predictive of the progression of RA (34). Taken together, these data imply that testing for anti-CCP antibodies allows the accurate identification of individuals who are at risk for the development of RA during the preclinical phase of the disease. More importantly, the identification of at-risk individuals would allow antigen-specific interventions before full-blown RA becomes established, thereby making it feasible to implement primary precautions in an antigen-specific manner.
We would like to thank Denis Greuet (INSERM U475) and the staff of the Leiden University Medical Centre animal facility for the animal care, Michèle Radal (Centre de Recherches et de Lutte contre le Cancer Val d'Aurelle, Montpellier, France) for histologic work, and Magali Terme, from the Laurence Zitvogel group (IGR, Paris, France) and Roald Pfannes for their helpful discussion concerning DC propagation in vitro.
- 27A comparative analysis of B cell-mediated myelin oligodendrocyte glycoprotein-experimental autoimmune encephalomyelitis pathogenesis in B cell-deficient mice reveals an effect on demyelination. Eur J Immunol 2002; 32: 1939–46., , , , , .