EAE, an animal model for multiple sclerosis, is a Th17- and Th1-cell-mediated auto-immune disease, but the mechanisms leading to priming of encephalitogenicTcells in autoimmune neuroinflammation are poorly understood. To investigate the role of dendritic cells (DCs) in the initiation of autoimmuneTh17- andTh1-cell responses andEAE, we used mice transgenic for a simian diphtheria toxin receptor (DTR) expressed under the control of the murineCD11c promoter (CD11c-DTRmice onC57BL/6 background).EAEwas induced by immunization with myelin oligodendrocyte glycoprotein (MOG) protein in CFA. DCs were depleted on the day before and 8 days afterMOG immunization. The mean clinicalEAEscore was only mildly reduced inDC-depleted mice when DCs were ablated beforeEAEinduction. The frequency of activatedTh cells was not altered, andMOG-inducedTh17 orTh1-cell responses were not altered, in the spleens ofDC-depleted mice. Similar results were obtained ifDCswere ablated the first 10 days afterMOGimmunization with repeatedDCdepletions. Unexpectedly, transient depletion of DCs did not affect priming or differentiation of MOG-inducedTh17 andTh1-cell responses or the incidence ofEAE. Thus, the mechansim of priming ofTh cells inEAEremains to be elucidated.
Dendritic cells (DCs) are key actors of adaptive immune responses against infections [1-3]. There are several DC subsets in mice which are characterized by differential expression of cell surface markers and their location; e.g. myeloid DCs (mDCs), plasmacytoid DCs (pDCs), dermal DCs, CD11b+ DCs, and CD11b− DCs [4, 5]. Ly6Chi monocytes are considered to be precursors of inflammatory DCs (inflDCs) which are recruited to site of inflammation . InflDCs express intermediate to high levels of CD11c and MHC class II (MHC II). mDCs are highly specialized in priming naïve T cells in vitro . In vivo depletion of murine CD11c+ mDCs by diphtheria toxin (DTx)-based transgenic systems has demonstrated an indispensible role for DCs during priming of CD8+ T-cell responses to viruses, intracellular bacteria and malaria parasites [1, 6]. Priming of Th1 responses and Th2 responses to parasites also depends on DCs [2, 7]. Furthermore, ablation of DCs leads to dissemination of Streptococcus pyogenes . In contrast, constitutive ablation of CD11c+ DCs leads to spontaneous fatal autoimmunity, high numbers of Th17 and Th1 cells and production of autoantibodies such as antinuclear Ab . This suggests that DC-mediated deletion of autoreactive single-positive thymocytes is important and failure leads to the observed pathology . Constitutive deletion of DCs in vivo in lupus-prone mice results in amelioration of disease, but DCs are not required for initial priming of autoimmune Th cells. Instead, DCs are important for functional differentiation and expansion of T cells .
Little is known about the role of mDCs in priming and de novo differentiation of autoimmune Th cells in the organ-specific autoimmune disease EAE, an animal model for human multiple sclerosis . We have previously demonstrated that myelin oligodendrocyte glycoprotein (MOG)-induced EAE is mediated by MyD88-dependent mechanisms . IL-6 expression by mDCs depended on MyD88 and was upregulated between 4 and 10 days after MOG immunization . Furthermore, depletion of pDC prior to MOG immunization ameliorated the clinical and histopathological signs of MOG-induced EAE compared with control mice .
To dissect the roles of mDCs during the first 10 days after EAE induction when priming and de novo differentiation of autoimmune Th cells occur , we used a transgenic mouse model of inducible, short-term in vivo ablation of CD11chi DCs — but not pDCs, macrophages, or B cells . CD11c-DTR (where DTR stands for diphtheria toxin receptor) mice carry a transgene encoding a DTR-GFP fusion protein under the control of a murine CD11c promoter . Our results demonstrate a minimal if any effect if mDCs are deleted prior or during the first 10 days after induction of EAE by MOG immunization.
Efficiency of DC depletion
CD11c-DTR mice on C57BL/6 genetic background were immunized with MOG protein in CFA and pertussis toxin to induce EAE. First, the efficiency of DC depletion was assessed after DTx injection of CD11c-DTR mice. An analysis of DC depletion in the skin, skin-draining inguinal LN and spleen was performed both before and after MOG immunization. All results are presented in Supporting Information Table 1 and the most relevant results are presented in Figure 1. Dermal Langerin− DCs were efficiently depleted for at least 4 days after DTx injection and subsequent MOG immunization (Fig. 1A and Supporting Information Table 1). CD11chi MHC II+ mDCs from skin-draining LNs and spleen were also efficiently depleted whereas around 50% of CD11cintermediate MHC II+ inflDCs were depleted by the DTx injection (Fig. 1B and C). Finally, the frequency of PDCA-1+ B220+ CD11clo pDCs was not affected by the DTx injection (data not included). Thus, dermal DCs and mDCs but not pDCs, were depleted by the DTx injection in CD11c-DTR mice, which is in concordance with previous studies .
EAE severity and immune reactivity in PBS-treated CD11c-DTR and DTx-treated C57BL/6 control mice
To test for any unspecific effects of DTx on EAE, DTx-treated C57BL/6 mice were included in all experiments. No differences between PBS-treated CD11c-DTR control mice and DTx-treated C57BL/6 control mice were observed in terms of EAE severity or observed immune reactivity (Table 1; and data not included). This suggests that DTx does not affect the clinical signs of EAE or immune reactivity toward MOG.
Table 1. Effect of DC depletion on clinical signs of MOG-induced EAE
Early DC depletion ameliorates the clinical signs of EAE
In EAE, DCs upregulate their IL-6 and IL-23/IL-12p40 expression, and primed and differentiated pathogenic Th cells can be detected 4–10 days after MOG immunization [12, 14]. To assess the role of DCs during inititation of EAE, DCs were depleted in vivo after MOG immunization. For inducible, short-term in vivo ablation of DCs, CD11c-DTR mice that carry a transgene encoding a DTR-GFP fusion protein under the control of the murine CD11c promoter were used. Conditional depletion is induced by injection of DTx, which leads to a 5- to 6-day ablation of DCs .
DCs were depleted in vivo on the day before — or 8 days after — EAE induction. DC depletion in CD11c-DTR mice by DTx injection 1 day before MOG immunization did not alter the incidence but reduced the mean maximum clinical EAE score compared with that of PBS-treated control CD11c-DTR mice (p = 0.05; Table 1; Fig. 2A) or DTx-injected C57BL/6 mice (Table 1). DCs depletion in CD11c-DTR mice 8 days after MOG immunization did not alter the incidence or the mean maximum EAE score compared with PBS-treated control CD11c-DTR mice (Table 1 and Fig. 2B). Thus, early depletion of DCs before MOG immunization only mildly reduced the disease severity but did not influence the incidence of EAE.
The frequency of Treg cells is not affected by DC depletion in CD11c-DTR mice
To examine the effect of DC depletion on FoxP3+ Treg cells, the Treg-cell numbers were assessed. DCs were depleted in vivo 1 day before MOG immunization and the frequency of absolute number of FoxP3+ CD3+ Treg cells per spleen was measured 10 days after MOG immunization by flow cytometry. The mean number of Treg cells per spleen did not differ between DC-depleted and control CD11c-DTR mice (Fig. 3). Thus, in contrast to constitutive DC ablation, short-time depletion of DCs does not appear to affect the Treg-cell responses in this system.
Generation of bone marrow chimeras
When the experiments described above were performed, low mortality of CD11c-DTR mice (one to two mice/experiment) was observed within the first week after DTx injection. In our hands, mortality increased over time when we ran new experiments (data not included), as described by others . Mortality was observed to the same extent in mice that had not received MOG injection, and the mortality was thus not caused by the MOG immunization (data not included), but probably due to aberrant DTR expression in nonimmune cells. To assure that immune cells were not depleted by the DTx injection, the frequency of B cells, CD11b+ cells, T cells and Ly6Chi CD11b+ monocytes were analyzed 24 h after DTx injection in the spleen from CD11c-DTR mice (Supporting Information Figure 1). The frequency of these cells was not affected by the DTx injection and EGFP expression was undetected in these cell types (data not included). Therefore, the increased mortality in CD11c-DTR mice was unlikely due to aberrant expression of DTR in immune cells other than mDCs.
To reduce the mortality in CD11c-DTR mice following DTx injection  and obtain a better experimental design, bone marrow chimeras were generated. Bone marrow from CD11c-DTR donors was injected into lethally irradiated C57BL/6 hosts 6 weeks before EAE induction. No mortality was observed in the bone marrow chimeras following DTx injection (data not included). The efficiency of the DC depletion was again assessed after DTx injection. Dermal DCs and mDCs from skin-draining LNs and spleen were depleted after DTx injection (Fig. 1D–F and Supporting Information Table 1). Similar to CD11c-DTR mice, around 50% of inflDC were depleted (Fig. 1E–F) but not pDCs (data not included). Depletion of mDCs and inflDCs in the CNS was analyzed at peak of EAE (day 13 after MOG immunization) when detectable amounts of DCs are present in the CNS . mDCs and inflDCs were abundant in both DC-depleted and controls and were as expected not depleted at this late time point (Fig. 1G). The inflDCs of the CNS expressed very high levels of CD11b (data not included). Thus, mDCs but not pDCs were depleted by the DTx injection in bone marrow chimeras to the same extent as in CD11c-DTR mice.
Severe clinical signs of EAE after DC depletion in bone marrow chimeras
DCs were depleted in vivo on the day before — or for the first 10 days after — EAE induction. DC depletion in bone marrow chimeras by DTx injection 1 day before MOG immunization did not alter the incidence or the mean maximum clinical EAE score compared with that of PBS-treated control bone marrow chimeras (Table 1 and Fig. 2C) or DTx-injected C57BL/6 mice (Table 1). DC depletion in bone marrow chimeras 1 day before, 3 and 6 days after MOG immunization did not alter the incidence or the mean maximum EAE score compared with PBS-treated control bone marrow chimeras (Table 1 and Fig. 2C). Thus, depletion of DCs before — or during the first 10 days after — MOG immunization in bone marrow chimeras did not influence the disease severity or the incidence of EAE.
The frequency of activated/memory Th cells is not affected by DC depletion
To assess the role of DCs during priming of autoimmune Th cells, DCs were depleted in vivo 1 day before MOG immunization in bone marrow chimeras. The frequency of naïve and act-ivated/memory Th cells were assessed 10 days after EAE induction by flow cytometry. Splenocytes were stained with Ab to CD62L, CD44, CD4, and CD3 and the frequency of naïve CD62Lhi CD44lo CD4+ T cells and activated/memory CD44hi CD4+ T cells was measured in DC-depleted or PBS-treated control MOG-immunized bone marrow chimeras and unimmunized mice (Fig. 4A). The mean frequency of activated/memory Th cells was much higher in both MOG-immunized groups compared with unimmunized mice (p < 0.004; Fig. 4B) and the mean frequency of naïve Th cells was much lower in both MOG-immunized groups compared with unimmunized mice (p < 0.004; Fig. 4B). The mean frequency of naïve or activated/memory CD4+ T cells did not however differ between MOG-immunized DC-depleted or control mice (Fig. 4B). The same results were obtained in mice that were treated with DTx 1 day before and 3 and 6 days after MOG immunization to deplete DCs for the entire period before analysis of Th-cell activation (data not included). This suggests that priming of encephalitogenic Th cells in vivo is not mediated by DCs, which is in concordance with data from a murine lupus model .
MOG-induced IL-17A-producing cells in DC-depleted bone marrow chimeras
To examine the effect of DC depletion on the Th17-cell responses, the absolute numbers of IL-17A-producing cells were measured by ELISPOT in the spleen 10 days after MOG immunization in bone marrow chimeras depleted of DCs in vivo 1 day before MOG immunization and subsequent restimulation with or without MOG ex vivo. Bone marrow chimeras treated with DTx 1 day before MOG immunization exhibited similar numbers of MOG-induced IL-17A-producing cells per spleen compared with PBS-treated control bone marrow chimeras (Fig. 5A). Both DC-depleted (p < 0.01) and PBS-treated controls (p < 0.05) exhibited however higher mean numbers of MOG-induced IL-17A-producing cells compared with unimmunized mice (Fig. 5A).
When DCs were depleted on day 5 after MOG immunization and mice were sacrificed 5 days later, no mice died from the DTx injection and therefore CD11c-DTR mice were used. The number of IL-17A-producing cells did not differ between CD11c-DTR mice treated with DTx 5 days after MOG immunization or control mice (Fig. 5B).
Th1-cell responses after DC depletion in bone marrow chimeras
To examine the effect of DC depletion on the Th1-cell responses to MOG, the absolute numbers of Th1 cells were measured in the spleen 10 days after MOG immunization in bone marrow chimeras. Mice were DTx- or PBS-treated 1 day before EAE induction. Both MOG-immunized groups exhibited higher numbers of Th1 cells compared with unimmunized mice (p < 0.05; Fig. 6A). MOG-immunized, DC-depleted mice displayed similar numbers of MOG-induced Th1 cells per spleen as did MOG-immunized, PBS-treated mice (Fig. 6A). The same results were observed in CD11c-DTR mice that were DC-depleted or PBS-treated 5 days after MOG immunization (Fig. 6B). Thus, the Th1-cell reactivity to MOG is not affected by the DC depletion.
T-cell responses toward the adjuvant component Mycobacterium tuberculosis after DC depletion
Next, we investigated whether the immune reactivity toward a component of CFA, heat-killed Mycobacterium tuberculosis (M.tb), was altered after DC depletion. DCs were depleted 1 day before MOG immunization in DTx- or PBS-injected bone marrow chimeras. Ten days after MOG immunization, splenocytes were stimulated for 48 h with or without killed M.tb. The number of M.tb-induced IL-17A-producing cells was a tenfold lower than MOG-induced IL-17A-producing cells and did not differ between DC-depleted and control mice (Fig. 5A). The strength of the Th1 response was lower to M.tb than to MOG, but did not differ between DC-depleted and control mice (Fig. 6A). Thus, it appears that the immune reactivity to M.tb is not affected by the DC depletion and the IL-17A-producing cell response to M.tb is much lower than to MOG.
It is generally believed that DCs are critical for priming and activation of naïve T cells . In addition, DCs play a prominent role in expansion of Treg cells . Most of the experimental evidence comes, however, from studies of monocyte-derived DCs pulsed with antigen in vitro  or targeting of Ag to molecules expressed on mDCs [17, 18]. Transgenic systems for transient or constituitve ablation of DCs in vivo have been developed during the last years. In vivo ablation of DCs reveals a more complex role for DCs than anticipated. It is clear that DCs control the adaptive immune response during bacterial, viral, and parasitic infections [2, 6-8]. In contrast, constitutive ablation of DCs results in spontanous fatal autoimmunity . To avoid spontanous autoimmunity, we used conditional ablation of DCs in actively induced EAE. The clinical signs of EAE were only mildly ameliorated if DCs were depleted a day before EAE induction, but not if DCs were depleted 8 days after immunization. In addition, DC-depleted bone marrow chimeras showed similar EAE scores as controls. The incidence of EAE was however not affected by DC depletion in our transient system.
In agreement with a recent study in murine lupus , DC ablation did not affect priming of the Th cells. The proprotion of activated/memory Th cells was much larger during initiation of EAE than in unimmunized mice — even in the absence of DCs — and was not affected by the nearly complete mDC depletion. If DCs were the primary APC for priming naïve Th cells in EAE, an increased naïve Th-cell compartment after DC depletion would be expected. Thus, our data argues for that another cell type is the primary APC for priming naïve Th cells to become autoimmune.
Differentiation of Th17 or Th1 cells was also not affected by the DC depletion. Since we have previously shown that pDCs regulate the Th17 response toward MOG in EAE , we tested whether pDCs were also depleted in CD11c-DTR and bone marrow chimeras after DTx treatment. Two different flow cytometry methods clearly showed that pDCs were not depleted by the DTx injection. To further examine the role of DCs on Th differentiation, DC maturation and Treg-cell responses were studied. DC maturation 10 days after MOG immunization was not impaired after DC ablation a day before EAE induction. We have previously shown that IL-6 and IL-23p40 expression is upregulated in mDCs by a MyD88-dependent mechanism in EAE . Another possiblity was that Treg cells were affected by the DC depletion and subsequently ameliorated the EAE severity. The number of Treg cells in the spleen was however not affected by the DC depletion. After constituitive ablation of DCs, Treg-cell numbers are reduced [9, 10]. The difference between our data and their systems is probably caused by the short ablation period and the fact that thymic selection prior to DTx injection is most likely not affected in our system.
Others have clearly demonstrated that DCs reactivate primed encephalitogenic Th cells in the CNS during development of EAE . In their system, the myelin-reactive Th cells were however transferred to the mice after priming. In an EAE model of epitope spreading, naïve Th cells reactive to proteolipid protein139–151 were primed probably by DCs in the CNS . An ongoing myelin-reactive Th-cell response was required for epitope spreading to occur. The infiltration of DCs into the CNS was not affected in our transient system, and we focused on priming and de novo differentiation of naïve Th cells to become myelin-reactive, where DCs appear to have no major role or are redundant.
A reduced or an abolished CD11c expression on DCs during the development of EAE could have rendered the CD11c-DTR mice and bone marrow chimeras resistant to the DC depletion and skewed our results. We have however previously observed similar numbers of CD11chi MHC II+ mDCs in the spleen during sorting of mDC at 4 and 10 days after MOG immunization and in unimmunized mice  (A. Lobell, unpublished observations). It is therefore unlikely that reduced CD11c expression explains the observed phenotype.
Unexpectedly, transient ablation of DCs before or after EAE induction does not affect priming of Th cells or de novo differentiation of autoimmune, MOG-induced Th17 and Th1-cell responses. It suggests that DCs play a minor or a redundant role during initiation of pathogenic Th-cell responses in EAE and the mechanism of priming Th cells in EAE remains to be elucidated.
Materials and methods
Escherichia coli-derived rat MOG1–125 was produced as previously described . MOG consists of aa 1–125 of the extracellular part of native MOG and a histidin tag at the C terminus.
For in vivo ablation of DCs, CD11c-DTR mice that carry a transgene encoding a simian DTR-GFP fusion protein under the control of the murine CD11c promoter were generated as described  and obtained from Jackson Laboratory (Bar Harbor, ME, USA). C57BL/6 female mice, obtained from Taconic (Denmark), were bred at the animal house at Rudbeck laboratories, Uppsala University. All animals were kept at specific pathogen-free conditions and all studies have been reviewed and approved by the local ethical committee and all experiments were carried out in accordance with EU Directive 2010/63/EU.
Generation of bone marrow chimeras
Femur and tibiae bones were removed from euthanized CD11c-DTR female mice. Bone marrow was flushed out with DMEM supplemented with 10% FCS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 292 μg/mL L-glutamine (DMEM complete) (all from Invitrogen, Carlsbad, CA, USA). Ten million bone marrow cells were injected i.v. into lethally irradiated (8 Gy) 6-week-old C57BL/6 female mice (Taconic). The bone marrow chimeras rested for 6 weeks before the experiments commenced.
Age and sex-matched 9- to 17-week-old female mice were immunized with 200–260 μg of MOG in CFA containing 0.5 mg M.tb H37RA (Difco, BD Diagnostic systems, Sparks, MD, USA) in IFA (Sigma-Aldrich, St. Louis, MO, USA) s.c. at the day of immunization and 2 days after, mice were injected with 200 ng of pertussis toxin (Sigma-Aldrich) in 200 μL PBS i.p. Clinical symptoms of EAE were scored daily as follows: 1, tail weakness or tail paralysis; 2, hind leg paraparesis; 3, partial hind leg paralysis; 4, complete hind leg paralysis; 5, tetraplegia, moribund state or death caused by EAE.
In vivo DC depletion
To deplete DC in vivo, CD11c-DTR mice or bone marrow chimeras were injected i.p. with 100 ng DTx (Sigma-Aldrich) in 100 μL as previously described . Injection of CD11c-DTR mice or bone marrow chimeras with the same amount of PBS served as a control. To determine the efficiency of the ablation, DCs in dermis (Langerin− CD11c+ MHC II+ or Langerin+), skin-draining inguinal LN (CD11chi MHC II+), and spleen (CD11chi MHC II+) from DTx-treated mice were measured by flow cytometry 24 h after DTx injection or 3, 10, or 13 days after MOG immunization. To test whether pDC were also depleted, CD11clo B220+ PDCA-1+ cells in the spleen from DTx-treated mice were measured by flow cytometry 24 h after DTx injection.
Spleens were harvested 10 days after MOG immunization or from unimmunized mice, cells were resuspended in DMEM (SVA, Uppsala, Sweden) and filtered through a 40 μm cellstrainer (Falcon BD). Splenocytes were cultured in DMEM complete with or without 5 μg/mL MOG or 5 μg/mL M.tb for 48 h at 37°C and 5% CO2.
Flow cytometry was performed as previously descibed  with the addition of gating single cells in a FSC-A versus FSC-H plot. The following antimouse antibodies were used for staining from BD Biosciences (San Jose, CA, USA): CD3-fluorescein isothiocyanate (FITC), CD4-PerCP, IFN-γ-PE, CD11c-PE, PDCA-1-PE, MHC II-FITC, CD80-Alexa647, CD86-Alexa647, CD11b-PerCP-Cy5.5, B220-PerCP, Langerin-allophycocyanin, Ly6C-FITC, and isotype controls. Flow cytometry analysis was performed on a FACS Canto II cytometer (BD Biosciences).
Isolation of dermal single cells
Isolation of dermal single cells was performed as previously described .
Isolation of immune cells from the CNS
Isolation of immune cells using Percoll gradient from spinal cords was performed as presiously described . In the present EAE model, almost all infiltrates are located in the spinal cord . Therefore, spinal cords from three mice per group were pooled to obtain detectable amounts of cells for flow cytometry.
To assess the number of IL-17A-secreting splenocytes from MOG immunized or unimmunized mice, an ELISPOT method was used as previously described .
Differences between mean daily EAE scores for individual mice, gene expression, and cytokine levels were analyzed with Mann–Whitney's U-test. p-values lower than 0.05 were considered significant. All analyses were performed using Graphpad Prism™ 4.0 software.
We would like to thank Dr. Dan Kaplan and Dr. Botond Igyarto, Minnesota University, for sharing their protocol of isolation of dermal DCs; and Dr. Jenny H. Martinsson, Uppsala University, for sharing her protocol of generation of bone marrow chimeras. We would also like to thank Rakan Naboulsi for excellent technical assistance. This work was supported by a grant from The Swedish Research Council, The Swedish Research Council Formas, Petrus and Augusta Hedlund's Foundation, Tornspiran foundation, The Hoff family (via The Swedish Brain Foundation), The Swedish Association for the Neurologically Disabled, Torsten and Ragnar Söderbergs Foundation, The Lars Hierta Memorial Foundation, and Magnus Bergvall's Foundation. No funding source had any involvement in study design, collection, analysis, or interpretation of the data. Further, no funding source had any involvement in writing or submitting the paper.
Conflict of interest
The authors declare no financial or commercial conflict of interest.