Paeoniflorin Ameliorates Experimental Autoimmune Encephalomyelitis via Inhibition of Dendritic Cell Function and Th17 Cell Differentiation

Paeoniflorin (PF) is a monoterpene glycoside and exhibits multiple effects, including anti-inflammation and immunoregulation. To date, the effect of PF on multiple sclerosis (MS) has not been investigated. In this study, we investigated the effect of PF in experimental autoimmune encephalomyelitis (EAE), an animal model for MS. After administered with PF, the onset and clinical symptoms of EAE mice were significantly ameliorated, and the number of Th17 cells infiltrated in central nervous system (CNS) and spleen was also dramatically decreased. Instead of inhibiting the differentiation of Th17 cells directly, PF influenced Th17 cells via suppressing the expression of costimulatory molecules and the production of interlukin-6 (IL-6) of dendritic cells (DCs) in vivo and in vitro, which may be attributable to the inhibition of IKK/NF-κB and JNK signaling pathway. When naïve CD4+ T cells were co-cultured with PF-treated dendritic cells under Th17-polarizing condition, the percentage of Th17 cells and the phosphorylation of STAT3 were decreased, as well as the mRNA levels of IL-17, RORα, and RORγt. Our study provided insights into the role of PF as a unique therapeutic agent for the treatment of multiple sclerosis and illustrated the underlying mechanism of PF from a new perspective.

Multiple sclerosis (MS) is an autoimmune disease characterized by chronic inflammation and demyelination in the central nervous system (CNS). Experimental autoimmune encephalomyelitis (EAE) is a well-established animal model for studying the underlying pathogenesis and developing new therapies for MS 1,2 . The molecular mechanisms of the disease are complex, and it is widely accepted that autoreactive Th cells are critically involved in the development of this disease 3,4 . Numerous reports have well studied the significant role of Th1 cells in EAE, however, abundant evidence demonstrate that Th17 cells are the main effector cells in the pathogenesis of EAE [5][6][7] .
Antigen-presenting cells (APCs) play a critical role in the initiation and development of adaptive immune responses. Dendritic cells (DCs), as a professional APC, are uniquely capable of presenting antigenic peptides to naïve T cells for clonal expansion [8][9][10] . Bone marrow-derived DCs (BMDCs) located at peripheral tissue are immature DCs (iDCs) that differentiate into mature DCs (mDCs) when stimulated with antigen. Then the latter migrate to lymphoid tissue and participate in immune response. The phenotypes of DCs at different differentiation stages are diverse, including MHCII and costimulatory molecules CD80, CD86, and CD40, which interact with corresponding ligands to provide signals for activating T cells 11,12 . In addition, pro-inflammatory cytokines IL-12 and IL-6 secreted by DCs are crucial in promoting the differentiation of Th1 cells and Th17 cells, respectively [13][14][15][16] . Given the important role they play in the immune system, DCs are a potential target for the treatment of immune-mediated inflammatory diseases.
As a mixture of various bioactive components, total glucosides of paeony (TGP) are extracted from the roots of Paeonia lactiflora and have been widely used as an anti-inflammatory remedy in the therapy for autoimmune PF treatment alleviated CNS pathological manifestations. Inflammatory cell infiltration and demyelination are the major pathological manifestations in the CNS of EAE mice. To further confirm the anti-inflammatory effect of PF, H&E and Luxol fast blue staining were performed to analyze the neuropathological changes of spinal cord tissues. Compared to control group, PF administration remarkably attenuated the degree of infiltrated inflammatory cells and demyelination in EAE mice (Fig. 1C). The number and percentage of infiltrated MNCs in CNS of PF-treated and control EAE mice was analyzed by flow cytometry. The expression of CD3, CD4, CD8, B220, and CD11b on MNCs purified from CNS was applied for distinguishing infiltrated T-cell, B-cell, and macrophages/microglia, respectively. PF treatment resulted in significantly reduced percentages of CD4 + and CD8 + T cells. More importantly, the absolute numbers of these cells were remarkably decreased in PF-treated EAE mice compared with that in control EAE mice (Fig. 1D). Taken together, these results indicate that PF obviously ameliorates the inflammatory reaction of EAE.

Amelioration of EAE by PF treatment was correlated with decreased Th17 cells. To investigate
the underlying mechanism for anti-inflammatory property of PF, we first examined the profile of CD4 + T cells in EAE mice. As illustrated in Fig. 2A, when stimulated with MOG  peptide ex vivo, the proliferation of CD4 + T cells purified from PF-treated EAE mice was markedly dampened. Given the crucial role of autoreactive pathogenic Th1 and Th17 cells in the pathogenesis of EAE, the Th cell subsets of MNCs derived from CNS and spleen of EAE mice were analyzed by flow cytometry. Specifically, the percentage of Th17 (CD4 + IL-17 + ) cells both in the spleen and CNS was significantly reduced after PF treatment, whereas the relative numbers of Th1 (CD4 + IFN-γ + ) and Treg (CD4 + CD25 + Foxp3 + ) cells were not influenced obviously (Fig. 2B). Consistently, RORγ t and RORα , key transcriptional factors of Th17 cells, were obviously decreased in CD4 + T cells derived from PF-treated EAE mice (Fig. 2C). Moreover, levels of IL-17 and IL-6 (Th17 cytokines) were remarkably decreased by PF treatment (Fig. 2D). These data suggest that PF ameliorates inflammation of EAE via decreasing the percentage of Th17 cells in vivo.
To further delineate whether PF suppresses inflammation through influencing Th17 cells directly, we observed the effect of PF on the differentiation of Th17 cells in vitro. As shown in Fig. 2E, no obvious changes were detected in Th17 cell differentiation at different concentrations of PF. The result indicates that PF do not directly work on Th17 cells.
The expression of costimulatory molecules, MHC II, and the secretion of IL-6 by DCs were inhibited by PF in vivo and in vitro. APCs are critically involved in the initiation of adaptive immunity.
However, only DCs could prime naïve T cells for initiating antigen-specific CD4 T cell immune responses. Accumulating evidence has demonstrated that control of DCs can influence the development of autoimmune and inflammatory diseases 24,25 . Skarica et al. reported a significant reduction of Th1 and Th17 immune responses in EAE by inhibiting APCs 26 . Thus, it is crucial to further study whether PF could modulate DC functions. To this end, we first analyzed the phenotype of CD11c + DCs. Splenocytes derived from PF-treated and control EAE mice were analyzed for the expression of costimulatory molecules and MHC II in CD11c + gate by flow cytometry. As shown in Fig. 3A, the expression levels of CD80, CD40, and MHC II were decreased by PF treatment, whereas the surface expression of CD86 was not altered. Then we detected the effect of PF on the production of cytokines from DCs. The real-time PCR results showed that PF inhibited the expression of IL-6 in CD11c + DCs, whereas IL-12p35 and IL-12/23p40 were not significantly affected (Fig. 3B). To verify the inhibiting effect of PF on DCs, bone marrow cells were induced for BMDC differentiation at different concentrations of PF in vitro. In consistent, the expression of surface markers was inhibited by PF in a dose-dependent manner (Fig. 3C). Moreover, IL-6 and IL-23 production of BMDCs was decreased by PF in vitro (Fig. 3D), which was consistent with the observed effect of PF in vivo. These results suggest that PF could suppress DC function and thus lead to ameliorating the inflammatory reaction of EAE.
PF impaired Th17 cell differentiation through inhibiting IL-6 production by DCs. It has been confirmed that cytokines derived from DCs play a controlling role in Th cell subset differentiation. Because the inhibition of PF on Th17 cell differentiation was only observed in vivo in this study, but not in vitro, we assumed that PF might affect Th17 cell polarization indirectly by regulating DCs-derived cytokines. Therefore, we examined whether the reduction of CD11c + DCs-derived IL-6 would affect Th17 cell differentiation. For this purpose, naïve CD4 + (CD4 + CD62L + ) T cells were co-cultured with CD11c + DCs sorted from PF-treated and control EAE mice under Th17-polarizing conditions to measure Th17 cell differentiation and the related events. IL-6 was known to play critical roles in Th17 cell differentiation 27 , and STAT3, which is activated by IL-6, is a key transcription factors for this differentiation 28 . As shown in our data, both the percentage of Th17 cells (Fig. 4A) and phosphorylation level of STAT3 (Fig. 4B) were markedly reduced in CD4 + T cells that cocultured with PF treatment CD11c + DCs. mRNA levels of IL-17, RORα and RORγ t were decreased consistently (Fig. 4C). Furthermore, the effect of PF-treated BMDCs on Th17 cell differentiation was measured by coculturing with the above naïve CD4 + T cells under Th17-polarizing conditions. As illustrated in Fig. 5A, the differentiation of Th17 cells was obviously impaired in PF dose-dependent manner. These data demonstrates that PF inhibits Th17 cell differentiation by decreasing the production of DCs-derived IL-6.

PF inhibited DC function via suppressing IKK/NF-κB and JNK activity. Previous studies have
confirmed the crucial role of IKK/NF-κ B activation in the function of DCs including costimulatory molecules expression and inflammatory cytokines production 29,30 . Once IKK/NF-κ B pathway is activated, Iκ B is activated by phosphorylated IKK and then degraded, followed by activated NF-κ B p65 translocating into the nucleus where it regulates target gene expression 31,32 . Therefore, the protein expression of IKK/NF-κ B signaling pathway and NF-κ B p65 nuclear translocation were detected by immunoblot and immunofluorescence assay, respectively. As present in our data, the protein expression of pho-Ikkα /β , pho-Iκ Bα and pho-NF-κ B p65 was reduced by PF treatment, and the degradation of Iκ Bα was also decreased (Fig. 5B). In accordance with the above results, the nuclear translocation of NF-κ B p65 was significantly decreased in PF-treated BMDCs (Fig. 5C). Since mitogen activated protein kinases (MAPKs) pathway plays important roles in inflammatory response 33 , the effects of PF treatment on MAPK pathway, including extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK) as well as p38 MAPK, were explored. Immunoblot data showed that the protein expression of p-JNK was substantially decreased with elevating in the concentration of PF, whereas the expression of p-ERK and p-p38 was not affected obviously (Fig. 5D). These results suggest that PF suppress DC function through inhibiting the activation of IKK/NF-κ B and JNK.

Discussion
Currently, it has been widely accepted that multiple elements of immune system and CNS play crucial roles in the pathogenesis and development of MS and EAE. The progression of this autoimmune disease covers numerous steps that can be divided into two stages: early phase including initiating and establishing the autoimmune response to myelin components in the periphery immune system, later phase associated with the destructive inflammatory responses in CNS including demyelination and axonal damage 34 . In the present study, we demonstrated that PF provided a significantly therapeutic effect on MOG 35-55 -induced EAE. As a result, PF treatment ameliorated the clinical signs, delayed the onset and reduced the incidence of EAE. The therapeutic effect of PF was associated with a significant reduction in inflammatory cell infiltration and the subsequent demyelination of CNS.
Autoreactive Th1 (producing IFN-γ ) and Th17 (producing IL-17) cells are primed and developed outside the CNS, once activated by myelin components, they enter into CNS and then further recruit other immune cells to activate resident macrophages/microglia, thus leading to the damage of CNS structures and resulting in progressive paralysis 2 . Adoptively transferring myelin-reactive Th1 and Th17 cells to naïve mice also can induce clinical signs of EAE. In addition, levels of IFN-γ and IL-17 are markedly elevated in MS patients and EAE mice 35 . Therefore, as drivers of autoimmune response, pathogenic Th1 and Th17 cells are considered as the principal criminal in MS and EAE 36 . Our data demonstrated that treatment with PF markedly decreased Th17 cells Values were expressed as mean ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001. percentage in the spleen and CNS at the peak period, whereas Th1 and Treg cells were not affected. Considering the result that PF delayed the onset of EAE, we proposed that PF may mainly inhibit the immune response in early stage, primarily through affecting Th17 cell polarization that has been thought as key factor in eliciting EAE 37 . However, inconsistent with the results in vivo, Th17 cells differentiation in vitro was not altered. These results suggested that PF suppressed Th17 cells differentiation via indirect means in vivo, resulting in the amelioration of EAE.
The activation of myelin-reactive Th cells is the predominant event in the progression of EAE. In addition, APCs also play an important role in the pathogenies and development of EAE. As professional APCs, DCs are uniquely able to directly activate naïve T cells and initiate autoimmune response in the periphery. Activation and differentiation of myelin-reactive Th cells are directed by DC through three signals. After being processed by DCs, antigen is presented to T cell recognition through MHCII, providing the first activation signal. The costimulatory molecules of CD80, CD86 and CD40 on the surface of DCs provide a second signal of T cell proliferation. The third signal is pro-inflammatory cytokines secreted by DCs, which determined the differentiation of activated Th cells 16 . Our data showed that PF significantly affected the early stage of immune response -Th cell activation. To further explore the mechanism of PF-induced amelioration of EAE, we next investigated the effect of PF on CD11c + DC in mice spleen in vivo and BMDCs in vitro which share a set of common features as naturally derived DCs 35 . Our results showed that PF reduced the expression of costimulatory molecule CD80 and CD40 as well as MHCII in vivo, in a dose dependent manner, whereas CD86 were not affected. Consistent with the data in vivo, the expression of CD80, CD40 and MHCII on BMDCs was significantly decreased. In this study, we demonstrated for the first time that PF decreased the CD40 expression in CD11c + DCs and BMDCs. A previous study has demonstrated the significantly decreased differentiation of Th17 rather than Th1 in CD40-dificient mice 38 , and another study detailed a necessary of CD40 expression on DC for inducing Th17 differentiation in vitro and in vivo 16 . Therefore, our results suggested that treatment with PF markedly decreased the presence of Th17 cells through suppressing the surface molecule expression on DCs.
Besides surface molecule expression on DCs, DC-derived cytokines are also predominant elements for inducing Th17 cell differentiation. Previous reports have described the predominant role of IL-6 and TGF-β in mouse Th17 cell induction [39][40][41] and IL-23 in maintaining and expanding the developing Th17 population 38,42 . However, TGF-β involves in both Th17 and Treg cell differentiation, promoting pro-inflammatory and anti-inflammatory response, respectively 43 . The presence of IL-6 seems to be a key determinant of Th17 polarization. IL-6 deficient mice are completely resistant to EAE induction 44 , and IL-6R blockade can prevent the development of EAE by inhibiting the number of MOG-active CD4 + T and CD8 + T cells in CNS and Th17 cells in peripheral lymphoid tissue 45 . However, neither pathogenic T cells nor CNS-resident cells are required to produce IL-6 in EAE. Instead, the requirement for IL-6 is entirely controlled by DC-derived IL-6 16,46 . These studies revealed that Th17 induction is critically dependent upon IL-6 derived from the initiated DCs. In this study, we found that PF treatment reduced IL-6 expression in CD11c + DCs from EAE mice in vivo and in BMDCs in vitro. Our results are consistent with other studies indicating that PF treatment reduced IL-6 production of macrophages 47,48 and decreased serum IL-6 levels in autoimmune hepatitis mice 49 . Moreover, we found that Th17 polarization was reduced when either CD11c + DCs from PF-treated EAE mice or PF-treated BMDCs were used as the source of IL-6. Our results indicated that PF inhibited IL-6 production of DCs, and thus resulted in decreased Th17 polarization.
The major role of IKK/NF-κ B signaling pathway in the pathogenesis of MS/EAE is well reported by various findings [50][51][52] . Previous studies showed that IKK/NF-κ B activation contributed to LPS-mediated surface molecule up-regulation and induced transcription of pro-inflammatory target genes 30 . Our study showed that PF suppressed the LPS mediated activation of IKK/NF-κ B signaling pathway in BMDCs by inhibiting the phosphorylation of IKKα /β , Iκ Bα and NF-κ B p65. Less amount of nuclear NF-κ B p65 in PF-treated BMDCs also proved the inhibition of PF on IKK/NF-κ B activation. In addition to the IKK/NF-kB pathway, LPS stimulation has been shown to activate MAPK signaling pathway. It has been demonstrated that three MAPKs, including ERK, JNK, and p38, are involved in the development of EAE [53][54][55] . In the present study, we illustrated that PF could significantly suppress the phosphorylation of JNK in BMDCs, suggesting that PF could suppress JNK cascades.
In summary, we demonstrated that PF treatment can effectively improve the clinical symptoms and delay the disease onset in EAE. In our present study, PF decreased IL-6 production and down-regulated costimulatory molecule expression in DCs via impairing the IKK/NF-κ B and JNK signal activation, thus leading to decreased inflammatory response induced by Th17 cells. Our findings support the potential therapeutic effect of PF for MS. Mice were observed daily and scored for disease severity on a scale of 0-5: 0, no clinical sign; 1, limp tail; 2, one hindlimb paralysis; 3, bilateral hindlimb paralysis; 4, hindlimb and forelimb paralysis; 5, moribund or dead. PF was administered i.p. at 100 μ g/mouse daily starting from 4d before immunization, and equal volume of PBS was served as control.

Mice and reagents.
Histopathological analysis. Spinal cords from PF-treated and control EAE mice were immediately immersed in 4% paraformaldehyde for fixation. After 2 days later, the specimen was embedded in paraffin for sectioning. The paraffin sections (5 μ m thickness) were stained with H&E and luxol fast blue for assessing the inflammatory cell infiltration and demyelination, respectively.

Isolation of Mononuclear cells. To isolate the infiltrating mononuclear cells (MNCs) from spinal cord and
brain (referred to as CNS hereafter), cardiac perfusion with PBS was first performed in EAE mice to eliminate the peripheral blood cells. The dissociated CNS tissue was gently grinded to prepare for cell suspension. MNCs from CNS were isolated using Percoll (GE Healthcare, 17-0891-02, MD, USA) density gradient (37% and 70%) centrifugation.

Isolation of DCs from mouse spleen.
For isolation of spleen DCs, spleens were cut into small pieces and incubated for 1 h at 37 °C with 1 mg/ml collagenase D (Roche, 11088866001, CA, USA) and 0.02 mg/ml DNase I (Roche, 11284932001) in RPMI-1640. Single cell suspension was prepared by grinding the small pieces through a 70 μ m cell strainer. Then cells were blocked by FcR Blocking Reagent (eBioscience, 14-0161-85, 93). CD11c + cells were magnetically sorted by CD11c MicroBeads (Miltenyi Biotech, 130-097-059) according to the manufacturer's instruction.
Scientific RepoRts | 7:41887 | DOI: 10.1038/srep41887 For co-culturing with BMDCs, the ratio was 20:1. After 3 days of culture, cells were stained with Abs against IL-17 and IFN-γ or p-STAT3, and then analyzed by flow cytometry as described above.
Immunofluorescence assay. BMDCs without or with PF (5 μ M) treatment were stimulated with LPS (100 ng/ml) for 30 min. Then cells were fixed with 4% paraformaldehyde for 20 min and further permeabilized with 0.1% TritonX-100 for 5 min. After washing with PBS, cells were collected for preparing cell smears. The smears were blocked with 1% bovine serum albumin for 45 min, treated with anti-NF-κ B-p65 antibody (1:400) at 4 °C overnight, followed by Alexa Fluor 555-labeled anti-rabbit IgG (1:1000; Cell Signaling Technology, 4413) incubation. DAPI was used to stain the nuclei. Slides were observed under a fluorescence microscope (Olympus, Tokyo, Japan).

Statistical analysis.
All experiments were repeated for at least three times and data were shown as mean ± SEM. One-way ANOVA was performed to determine whether an overall statistically significant change existed before the Student's t-test was used to analyze the difference between two groups. Differences were considered significant when the P value was < 0.05.