Cordycepin Prevents and Ameliorates Experimental Autoimmune Encephalomyelitis by Inhibiting Leukocyte Inltration and Reducing Neuroinammation

Multiple sclerosis (MS) is a neuroinammatory autoimmune disease characterized by multifocal perivascular inltration of immune cells in the central nervous system (CNS). Current treatment for MS is unsatisfactory, and we aimed to search for immunomodulatory agents from bioactive constituents of natural origin. Cordycepin (3'-deoxyadenosine), an adenosine analogue initially extracted from the fungus Cordyceps militarisa, is one of the candidates that has multiple actions. However, its effect on MS is unknown. Human MoDCs from three donors were cotreated with LPS (0.1 μg/mL) combined with 50, 100 and 150 μg/mL cordycepin (CO50, CO100 and CO150) for 18 h. The levels of TNF-α and IL-6 in the culture supernatant were measured using ELISAs. The results are presented as the means + SD. (D) BMDCs were cotreated with LPS (0.1 μg/mL) in the presence of 0, 12.5, 25 and 50 μg/mL cordycepin (CO) for 30 min. Levels of AKT/p-AKT, ERK/p-ERK, NFκB/p-NFκB, and actin were determined using Western blotting. (E) BMDCs were pretreated with 25, 50 and 100 μg/mL cordycepin (CO25, CO50 and CO100) for 1 h and then stimulated with/without 0.1 μg/mL LPS for 3 h. ROS generation within the cells was determined by measuring the absorbance at Ex/Em wavelengths of 495/529 nm. An ROS inducer was used as an experimental control according to the manufacturer’s protocol.


Introduction
Human MoDCs were harvested as described previously [25]. Human peripheral blood mononuclear cells (PBMCs) were isolated from human blood using Ficoll and Percoll gradient centrifugation. Monocytes were enriched by adherence as follows: 8 × 10 5 PBMCs per milliliter were seeded in 10 cm dishes containing 10 ml of complete RPMI 1640 medium, incubated for 18 h, and then the suspended cells were removed by washing with RPMI 1640 medium. After washing, monocytes were cultured in 10 ml of complete RPMI 1640 medium supplemented with 100 ng/ml human GM-CSF (PeproTech Inc., New Jersey, USA) and 100 ng/ml human IL-4 (PeproTech Inc., New Jersey, USA). Another 10 ml of complete RPMI 1640 medium containing 100 ng/ml GM-CSF and IL-4 was added on day 3. The MoDCs were collected from each dish, washed, and counted on day 6. The study involving human subjects was approved by the Research Ethics Committee of China Medical University & Hospital, Taiwan (CMUH110-REC2-052).

Measurement of surface markers and cytokines from dendritic cells and macrophages
Mouse BMDCs, human MoDCs or RAW264.7 macrophages were plated at a density of 1 × 10 6 cells per milliliter in complete RPMI 1640 medium. Then, 0.1 μg/ml lipopolysaccharide (LPS) (Sigma, MO, USA) with or without the indicated concentration of cordycepin (Sigma, USA) were added, and the cells were subsequently incubated for 18 h at 37°C in a 5% CO 2 atmosphere. The cells were incubated with 0.1 μg/ml LPS as a stimulator. After the incubation, the cells were harvested, and uorescence-labeled mouse/human anti-CD11c, anti-CD40, anti-CD86, anti-mouse I-A b (MHC II), anti-CCR7 (BD Bioscience, CA, USA), anti-integrin b1, anti-integrin a4, anti-LFA-1, anti-c type lectin and anti-ICAM-1 (Biolegend, CA, USA) monoclonal antibodies were used to stain DC surface markers. The expression of these markers was analyzed using a FACSVerse instrument (BD Bioscience, CA, USA). The supernatants from cells cultured for 18 h were isolated and assayed for mouse/human IL-6, TNF-α and IL-12 levels using ELISA kits (BD Bioscience, CA, USA) according to the manufacturer's protocol.

Western Blotting
Mouse BMDCs (1 × 10 6 cells/ml) were plated in complete RPMI 1640 medium. Then, 0.1 μg/ml LPS (Sigma, MO, USA) with or without the indicated concentration of cordycepin were added, and the cells were further incubated for 30 min at 37°C in a 5% CO 2 atmosphere. Whole-cell lysates (15 μg per lane) were separated by electrophoresis on 10% sodium dodecyl sulfate-polyacrylamide gels and then analyzed by Western blotting using speci c antibodies against AKT/p-AKT, ERK/p-ERK, NFκB/p-NFκB (Cell Signaling Technology, MA, USA), and actin (Proteintech, China). Actin served as a loading control.

Measurement of Reactive Oxygen Species (ROS)
BMDCs (5 × 10 5 /ml) were cultured in the presence of ddH 2 O (control) or the indicated concentration of cordycepin for 1 h and then stimulated with LPS (0.1 μg/ml) for 3 h to detect the intracellular ROS level. After stimulation, the media were replaced with fresh medium containing ROS Label buffer (BioVision, CA, USA) and incubated for 20 min at 37°C before thorough washing with PBS. The intracellular ROS level was assayed using an ROS Detection Assay Kit (BioVision, CA, USA) according to the manufacturer's protocol. The uorescent signal was detected at Ex/Em= 495/529 nm in end point mode in the presence of compounds and controls.
In vitro migration assay BMDCs (1 × 10 6 ) were treated with 0.1 μg/ml LPS in the presence or absence of 50 μg/ml cordycepin for 18 h prior to the migration studies. After washing and counting, 2 × 10 5 BMDCs from different groups in 100 μl were transferred to the upper chamber (Millicell cell culture inserts, pore size: 5.0 μm) (Millipore, Germany) of a 24-well plate. The upper chamber contained cells in 100 µl media lacking cytokines and serum. The lower chamber contained 500 µl of complete media (10% FBS) treated with or without CCL21 (250 ng/ml). The upper chamber was removed after 6 h, and the cells that migrated into the lower chamber were counted using ow cytometry and WST-1 staining (Takara Bio, CA, USA).
Experimental autoimmune encephalomyelitis (EAE) model C57BL/6 mice were obtained from the National Laboratory Animal Breeding and Research Center, Taipei, Taiwan. All mice were housed in a speci c pathogen-free room at the Animal Center, China Medical University, Taichung, Taiwan and maintained in accordance with institutional animal care guidelines. The animal protocol was approved by the Animal Care and Use Committee of China Medical University, Taichung, Taiwan (CMUIACUC-2018-144).
Eight-to ten-week-old female C57BL/6 mice were immunized subcutaneously (s.c.) with 200 μg of myelin peptide (MOG  ) (Kelowna International Scienti c INC, Taiwan) and 600 μg of mycobacterium tuberculosis (BD Bioscience, MD, USA) emulsi ed in 100 μl of PBS and 100 μl of complete Freund's adjuvant (CFA) (BD Bioscience, MD, USA) per mouse, as described previously [35]. Pertussis toxin (PTX; 400 ng; List Biological Laboratories, CA, USA) was administered intraperitoneally (i.p.) on the day of immunization and on day 2. Mice were monitored daily and scored for EAE as follows: 0, no clinical signs; 0.5, partially limp tail; 1, paralyzed tail; 2, hind limb paresis; 2.5, one hind limb paralyzed; 3, both hind limbs paralyzed; 3.5, weakness in forelimbs; 4, forelimbs paralyzed; and 5, moribund or dead animals [35]. Each experimental group contained 4-5 mice. Wild-type (WT; n=3) mice were not administered any treatment, while mice in the control group (CFA; n=3-5) were similarly s.c. immunized with CFA and PTX without MOG. Mean clinical scores on separate days were calculated by adding scores of individual mice and dividing by the number of mice in each group. The statistical analysis was performed using Student's t-test.
Preventive effect of cordycepin treatment on the EAE model Animals were i.p. injected with 50 mg/kg cordycepin (EAE+CO50; n=6) or saline (EAE+saline; n=6) daily from days 0 to 21 after immunization to assess the preventative effect of cordycepin on EAE. The CFA group (CFA; n=3) was similarly immunized without MOG. Disease scores were recorded daily as mentioned above. On the day of sacri ce (day 21 after immunization), the brains and spinal cords were harvested from anesthetized mice after an intracardiac infusion with PBS. Subsequently, the brains and spinal cords were xed with 4% paraformaldehyde. Fixed samples were embedded in para n, and cross sections were stained with hematoxylin-eosin (H&E) (service of Rapid Science, Taiwan) to analyze in ammatory cell in ltration.
We also established other experiments to examine the percentages of dendritic cells and T cells in the brains and spleens on day 8 after immunization. Brains and spleens were harvested and stained with a uorophore-labeled mouse anti-CD45 Ab, anti-CD11c Ab, anti-CD4 Ab, and anti-CD8 Ab (BD Bioscience, CA, USA) and analyzed using ow cytometry. Additionally, splenocytes were restimulated with MOG peptide ex vivo to determine the capacity of cytokine production. A total of 1 × 10 6 spleen cells were cultured in 24-well culture plates in the presence of 10 μg/ml MOG peptide for 3-4 days. The supernatant of cultured cells was collected and assayed for IL-6, TNF-α, IL-17A, and IFN-γ levels (BD Bioscience, CA, USA) using ELISA kits according to the manufacturer's protocol.
Therapeutic effect of cordycepin treatment on the early disease onset stage and late disease progression stage Cordycepin (n=5) or saline as a vehicle control (n=5) was i.p. administered daily from day 11 to day 22 to evaluate its effect on the early disease onset stage. Cordycepin (n=5), the positive control ngolimod (FTY720; n=5) (Sigma, USA), the vehicle control saline (n=4) and the negative control CFA (n=5) were i.p. administered from day 24 to day 38 to evaluate the effect of cordycepin on the late disease progression stage. Disease scores were recorded daily. Brains were harvested to detect IFN-γ+ CD4+ T cell and IL-17A+ CD4+ T cell populations. Cell surface markers ( uorochrome-conjugated anti-CD45 Ab and anti-CD4 Ab) (BD Biosciences, CA, USA) were labeled rst, and intracellular proteins were stained with uorochrome-conjugated anti-IFN-γ Ab and anti-IL-17 Ab (BD Biosciences, CA, USA) after xation and permeabilization. The populations of IFN-γ-and IL-17-producing CD4+ T cells in the brain were analyzed using ow cytometry.

Chemokine array
For the in vivo evaluation of chemokine production, the brains and spinal cords were harvested from different groups of mice on day 14 after immunization and intracardially perfused with PBS. Blood samples were harvested from mice by cardiac puncture and centrifuged at 12,800 ×g for 10 min to collect serum. Proteins were isolated by homogenizing tissues in RIPA buffer (Abcam, Cambridge, UK). Chemokine production was assessed using a Mouse Chemokine Array Kit (R&D Systems, MN, USA) according to the manufacturer's instructions. Array images were analyzed by densitometry using Image Lab Software (Bio-Rad). The data are presented as fold changes in chemokine expression compared to the corresponding positive control spots.

Next-generation sequencing (NGS)
The spinal cords were harvested from cordycepin-treated and untreated EAE mice on day 17 after immunization, and total RNA was extracted from the spinal cord using an RNeasy Micro Kit (Qiagen, Germany). RNA quantity and quality were assessed with a Bioanalyzer 2100 device using the RNA 6000 Nano Kit (Agilent Technologies). Differential gene expression analyses were conducted using RNA-seq quanti cation according to the Illumina procedure and were performed by Genomics Ltd. (New Taipei City, Taiwan). Differentially expressed genes (DEGs) were calculated using EBSeq, and functional analyses, including KEGG analyses, were performed using clusterPro ler. RNA-seq was performed using the Illumina platform according to the manufacturer's protocols. DEGs were calculated using EBSeq, and functional analyses, including KEGG analyses, were analyzed using clusterPro ler. Both RNA-seq and subsequent analyses were performed by Genomics Ltd. (New Taipei City, Taiwan).

Preparation and stimulation of EOC13.31 microglial cultures
Both LADMAC and EOC13.31 cells were purchased from the Bioresource Collection and Research Center (BCRC; Taiwan). LADMAC cells were cultured in complete Eagle's Minimum Essential Medium (Gibco, NY, USA) supplemented with 10% FBS, and the supernatants were collected to produce conditioned medium. Microglial EOC13.31 cells were cultured in Dulbecco's modi ed Eagle's medium (DMEM; Gibco) containing 20% LADMAC conditioned medium in the presence of 5% CO 2 at 37°C. For the quanti cation of microglial responses, EOC13.31 cells were seeded in a 24-well plate at a density of 1 × 10 6 cells/ml. Then, the cells were stimulated with IFN-γ (PeproTech, Taiwan) and treated with or without cordycepin. The supernatants from cultured cells were collected after 24 h and assayed for TNF-α levels using an ELISA kit (BD Bioscience, CA, USA) according to the manufacturer's protocol.

Results
Cordycepin inhibits antigen-presenting cell activation in a dose-dependent manner DCs are the most potent antigen-presenting cells for naive T cells. Immature DCs residing in the periphery have a strong ability to endocytose antigens and become mature DCs upon exposure to a variety of stimuli [36]. Some studies have investigated whether Chinese herbal medicines inhibit DC maturation to ameliorate autoimmune diseases and in ammatory responses [37][38][39][40][41]. First, we investigated the effect of cordycepin treatment on LPS-stimulated BMDCs. After 18 h, the levels of TNF-α, IL-6 and IL-12 produced by BMDCs were decreased (Fig. 1A). The expression of the costimulatory molecules CD40 and CD86 was also reduced in LPS-stimulated BMDCs in a dose-dependent manner (Fig. 1B). Furthermore, we examined the effect of cordycepin on activated human MoDCs. Monocytes from three donors were enriched and cultured to induce the formation of MoDCs. The secretion of TNF-α and IL-6 by LPS-stimulated human MoDCs was decreased following cordycepin treatment in a dose-dependent manner (Fig. 1C). Additionally, we determined the capacity of BMDCs to capture exogenous antigens after cordycepin treatment, and the uptake of FITC-labeled dextran by BMDCs was examined using ow cytometry. The internalization of FITC-labeled dextran in BMDCs was not signi cantly different after treatment with various concentrations of cordycepin (Supplementary Figure 1). Moreover, the MAPK, NF-κB and Akt signaling pathways and reactive oxygen species (ROS) have been shown to participate in the activation of DCs [42,43]. The levels of AKT, ERK, and NF-κB phosphorylation in LPS-stimulated BMDCs were reduced by a 30 min cordycepin treatment in a dose-dependent manner (Fig. 1D). ROS levels were decreased in LPS-induced activated BMDCs after the cordycepin pretreatment (Fig. 1E). Therefore, the maturation of LPS-stimulated DCs was suppressed by cordycepin treatment through the inhibition of the AKT, ERK, and NF-κB signaling pathways and ROS production.
Cordycepin downregulates adhesion molecules and chemokine receptors and impairs cell migration Activated DCs migrate from peripheral tissues to secondary lymphoid organs, which are probably highly arranged by a group of cell surface receptors and adhesion molecules [44]. According to previous studies, adhesion molecules, such as ICAM-1, c-type lectin, and integrin, expressed on DCs modulate cell crossing of the blood-brain barrier (BBB) [18][19][20]. We identi ed that cordycepin treatment suppressed the expression of adhesion molecules (integrin b1, integrin a4, c-type lectin, and ICAM-1) on LPS-stimulated BMDCs ( Fig. 2A). Moreover, CCR2, CCR5 and CCR7 are possible mediators of DC recruitment to the CNS in the EAE model and patients with MS [14][15][16][17]. Cordycepin treatment reduced the percentage of CCR7+ cells and the uorescence intensity of CCR7 but not CCR2 in LPS-stimulated BMDCs (Fig. 2B). However, we did not determine whether cordycepin treatment inhibited CCR5 production because LPS does not stimulate CCR5 expression in BMDCs (data not shown). Based on the aforementioned results, DC migration was tested in vitro using a transwell migration assay. The lower chamber was loaded with CCL21 (CCR7 ligand), and the upper chamber was loaded with BMDCs to assess migration toward CCL21. Pretreatment with LPS for 18 h increased BMDC migration, which was inhibited by cordycepin cotreatment, as measured using ow cytometry and the WST-1 assay (Fig. 2C). These results suggested that cordycepin treatment impairs dendritic cell migration by downregulating adhesion molecules and chemokine receptors.
Cordycepin ameliorates the clinical severity of EAE in the preventive model According to the results of the in vitro experiments described above, we further investigated whether cordycepin represented a potential immunosuppressive agent in mice with EAE, the murine model of MS. In the preventive model, cordycepin was administered daily beginning on day 0. Cordycepin treatment signi cantly ameliorated the clinical disease severity (Fig. 3A). Furthermore, we analyzed the level of in ltrated cells in the spinal cord by performing H&E staining on day 21. More in ltrated cells were detected in the spinal cord of EAE mice than in normal mice. Moreover, cordycepin treatment inhibited cell in ltration in the spinal cord of EAE mice (Fig. 3B). Additionally, we established another set of experiments to examine the populations of immune cells in the brains of EAE mice following cordycepin treatment on day 8. DCs, CD4+ T cells and CD8+ T cells in ltrated the CNS of EAE mice and cordycepin treatment blocked EAE-induced DC, CD4+ T cell and CD8+ T cell in ltration (Fig. 3C). In addition, fewer DCs, CD4+ T cells and CD8+ T cells remained in the spleen of EAE mice, and cordycepin treatment signi cantly reversed changes in the levels of DCs and CD8+ T cells (Fig. 3D). These data suggested that cordycepin treatment blocked immune cell in ltration into local CNS lesions; therefore, more immune cells remained in the systemic immune organ. On the other hand, we determined the speci c immune response of the spleen, a systemic immune organ, to the MOG antigen. We used a speci c peptide (MOG) to restimulate splenocytes derived from cordycepin-treated or untreated EAE mice, and then the speci c peptide-induced production of in ammatory cytokines, such as IFN-g (Th1 and Tc1-associated cytokines), IL-6, TNF-a (Th2-associated cytokines), and IL-17 (Th17-associated cytokines), were determined using ELISAs. Cordycepin treatment decreased the secretion of IL-6, TNF-a, IL-17 and IFN-g, and the levels of IL-6, TNF-a and IFN-g were particularly signi cantly reduced (Fig. 3E). Based on these results, cordycepin treatment exerts a preventive effect by inhibiting the in ltration of immune cells and may exert a therapeutic effect by attenuating the production of in ammatory cytokines.
Cordycepin inhibits the chemotactic response in the CNS Chemokines and their receptors are important mediators of the recruitment of immune cells to in ammatory sites. Chemokines are expressed at higher levels in patients with MS and recruit in ammatory cells into the CNS [45]. We used a chemokine array to analyze the changes in tra cking factor and chemokine protein expression in the brain, spinal cord and serum between EAE mice treated with or without cordycepin. Higher CCL6, PARRES2, IL-16, CXCL10, and CCL12 expression was detected in the brain and spinal cord of EAE mice than in normal mice (CFA group). Simultaneously, the high levels of CCL6, PARRES2, IL-16, CXCL10, and CCL12 were reversed in the brain and spinal cord of cordycepintreated EAE mice ( Fig. 4A and 4B). However, these changes were not observed in the serum (Fig. 4C). The fold changes in the levels of the indicated chemokines in the lower panel of Fig. 4A, 4B and 4C showed that cordycepin treatment inhibited CCL6 and CXCL10 production by more than ve-fold in the brain and spinal cord. The fold change in total chemokine expression is shown in Supplementary Figure 2. Furthermore, we examined the levels of adhesion molecules (CCR2, CCR5 and CCR7) expressed on in ltrated lymphocytes in the CNS ex vivo. Higher levels of adhesion molecules were detected on in ltrated lymphocytes in the spinal cord of EAE mice than in normal mice (CFA group). Moreover, lower levels of adhesion molecules were expressed on in ltrating lymphocytes in the spinal cord of cordycepintreated EAE mice than saline-treated EAE mice (Fig. 4D). We next assessed the transcriptomes of the spinal cord in response to cordycepin treatment in EAE mice. We con rmed that cordycepin treatment inhibited the expression of most adhesion molecules, chemokines, and chemokine receptors in the spinal cord of EAE mice (Fig. 5A). The gene expression levels of adhesion molecules (Fig. 5B), chemokine receptors (Fig. 5C), and chemokines (Fig. 5D) were consistent with the protein expression levels measured in vitro (Fig. 2) and in vivo (Fig. 4). The mRNA and protein expression levels of adhesion molecules, chemokine receptors, and chemokines were inhibited by cordycepin treatment in BMDCs or the CNS of EAE mice. These results suggested that cordycepin treatment reduces chemotactic activity in the CNS.

Cordycepin inhibits neuroin ammation and ameliorates the clinical severity of EAE in a therapeutic model
Macrophages and microglia promote neuroin ammatory and neurodegenerative changes in individuals with MS by releasing in ammatory cytokines and stimulating leukocyte activity and in ltration into the CNS [46]. To investigate the effect of activated microglia on cordycepin treatment, We stimulated mouse EOC13.31 microglial cells with IFN-γ, which mimics the in ammatory environment in vivo, to investigate the effect of activated microglia on the changes induced by cordycepin treatment. The increased level of TNF-α induced by IFN-γ stimulation was signi cantly reduced by cordycepin treatment in a concentrationdependent manner (Fig. 6A). We used LPS to stimulate the murine macrophage-like cell line RAW264.7, and the levels of TNF-α and IL-6 were determined using ELISAs to con rm that cordycepin inhibited in ammatory cytokine production from macrophages. Cordycepin treatment decreased the secretion of TNF-α and IL-6 by LPS-stimulated macrophages in a dose-dependent manner (Fig. 6B). These in vitro experiments suggested that cordycepin has the potential to inhibit neuroin ammation in individuals with MS. Moreover, IL-17-producing CD4+ T cells (Th17), IFN-γ-producing CD8+ T cells (Tc1), and IFN-γproducing CD4+ T cells (Th1) are known to play important roles in the progression of EAE and MS [6,8]. We treated Th17, Tc1, and Th1 cells with cordycepin in vitro to analyze whether it affected proin ammatory cytokine production. Cordycepin treatment reduced the percentages of activated IL-17producing Th17 cells, IFN-γ-producing Tc1 cells, and IFN-γ-producing Th1 cells (Fig. 6C). Furthermore, we determined whether cordycepin treatment exerted a therapeutic effect on the model of CNS in ammation (morbidity). In the early disease onset stage, cordycepin was administered daily beginning on day 11 after immunization, and the therapeutic effect was statistically signi cant throughout the period from day 14 to day 18 (Fig. 6D). In the late disease progression stage, cordycepin or ngolimod was administered daily beginning on day 24 after immunization, and both compounds exerted signi cant therapeutic effects ≥ day 35 after immunization (Fig. 6E). Furthermore, cordycepin and ngolimod treatment reduced the percentages of IL-17-producing T cells and IFN-γ-producing T cells in the brain in the late disease progression stage (Fig. 6F). Therefore, cordycepin treatment not only exerts a preventive effect by blocking immune cell in ltration but also exerts a therapeutic effect by reducing neuroin ammation to treat multiple sclerosis.

Discussion
This study sought to determine the preventive and therapeutic effects of cordycepin and its possible underlying mechanism in regulating immune cell in ltration and neuroin ammation. The small compound cordycepin from Cordyceps militaris inhibited LPS-induced dendritic cell activation by inhibiting the AKT, ERK, and NF-κB signaling pathways and ROS production (Fig. 1), impaired dendritic cell migration through the downregulation of adhesion molecules and chemokine receptors (Fig. 2), and blocked recruitment of immune cells to the CNS (Fig. 3) by decreasing the expression of tra cking factors in the CNS (Fig. 4-5). These processes may be the mechanism by which cordycepin treatment attenuated the clinical symptoms in the preventive EAE model. Moreover, cordycepin treatment inhibited the production of proin ammatory cytokines, such as IFN-g , IL-6, TNF-a, and IL-17, in EAE mice treated with a speci c peptide (Fig. 3E). We further identi ed that cordycepin treatment decreased cytokine production in activated microglial cells, macrophages, Th17 cells, Tc1 cells, and Th1 cells (Fig. 6). The inhibition of systemic in ammation and local neuroin ammation may be the mechanism by which cordycepin treatment attenuated the clinical symptoms in the therapeutic EAE model (Fig. 6). Fig. 7 shows a schematic diagram of the possible mechanism of cordycepin treatment in EAE mice.
The early hallmarks of MS pathogenesis are immune cell tra cking into the CNS and degradation of the endothelial BBB. The migration of immune cells to the CNS is a multistep process regulated by the sequential interaction of different adhesion molecules on the BBB and on immune cells. Brie y, circulating T cells roll on the endothelial surface via α4β1-integrin and LFA-1 engagement of endothelial VCAM-1, ICAM-1, and ICAM-2 [47,48]. Some molecules were previously shown to participate in the recruitment of various subsets of leucocytes to the CNS, including CD4+ T cells (LFA-1 and ALCAM) [49,50], CD8+ T cells (α4 integrin) [51], Th17 cells (CCR6 and CCL20) [52], and monocytes/dendritic cells (αL integrin, ALCAM, CCR2, CCR5, and CCR7) [14-17, 50, 53]. In addition, several proin ammatory cytokines and chemokines play a role in MS pathogenesis by disrupting the BBB and increasing immune cell tra cking, such as TNF-α, IFN-γ, IL-1, IL-6, IL-16, IL-17, IL-22, MMPs, CXCL9, CXCL10, and CXCL11 [54][55][56][57][58]. In the current treatment of multiple sclerosis, some drugs are mainly used to suppress the immune response. Fingolimod is an agonist of sphingosine-1 phosphate receptors 1, 3, 4, and 5 that sequesters lymphocytes in lymph nodes by inhibiting their egress into lymph [59], reduces the death of human brain microvascular endothelial cells induced by high concentrations of proin ammatory cytokines and downregulates vascular adhesion molecules such as ICAM-1, VCAM-1, and P-selectin on the BBB endothelium [60,61]. Laquinimod is a derivative of the immunomodulator linomide (quinoline-3carboxamide) that reduces ICAM-1 and ALCAM mRNA expression in the brain endothelium and blocks the migration of Th1 and Th17 cells across brain endothelial cells [62]. Natalizumab is a humanized monoclonal antibody that selectively blocks the VLA-4-mediated interaction of autoaggressive T cells with endothelial VCAM-1 on the BBB and inhibits T cell entry into the CNS in patients [63]. These clinical medicines targeting immune cell tra cking have proven to be a successful therapy for MS but have side effects and are no longer effective once patients have entered the progressive phase of the disease.
In the present study, cordycepin treatment reduce levels of the integrin b1, integrin a4, c-type lectin, ICAM-1, and CCR7 proteins on activated DCs and impaired their migration in vitro (Fig. 2C) and in vivo (Fig. 3C). Our chemokine array indicated that cordycepin treatment reduced the protein levels of chemokines (CCL6, PARRES2, IL-16, CXCL10, and CCL12) expressed in the brain and spinal cord (Fig. 4) and blocked lymphocyte in ltration into the CNS of EAE mice ( Fig. 3B and 3C). In addition to the abovementioned tra cking factors, NGS data showed that cordycepin treatment also signi cantly inhibited the mRNA expression of other adhesion molecules, cytokines, and their receptors, such as CxCl14, Xcl1, Ccl24, Ccr9, Cxcr6, Itgb7, Cxcl2, Pecam1, Pvr, Selplg, CxCl12, Ccl4, Cd2, Cxcr3, Pvrl2, Cd226, Cxcr4, and Cx3cr1. Some of these tra cking factors have not been con rmed to be related to multiple sclerosis. Therefore, we will further study the roles of these factors in multiple sclerosis in the future.
Neuroin ammation is one cause of brain injury and involves glial activation and the release of in ammatory mediators, such as cytokines and chemokines. It is considered an event that induces neuronal dysfunction and the progression of neurodegenerative diseases such as MS, Alzheimer's disease and Parkinson's disease. Microglia are macrophage-like cells in the CNS that are activated and secrete various in ammatory cytokines. For example, IL-6 and TNF-α have been identi ed as the major drug targets for neurodegeneration [64]. Neuroin ammation in MS is caused by lymphocytes invading the central nervous system, accompanied by damage to BBB function and a strong glial response [65]. In the present study, cordycepin treatment reduced cytokine production from not only microglial cells and macrophages but also Th1, Th17, and Tc1 cells in vitro (Fig. 6A-C). Either in the early disease onset or late disease progression stage, lymphocytes had already in ltrated into the CNS, and cordycepin treatment still exerted a therapeutic effect (Fig. 6D-E). The therapeutic effect of cordycepin on these stages may be due to the reduction in neuroin ammation.
Cordycepin from Cordyceps militaris has been reported to exhibit anti-in ammatory, antioxidant [66] and neuroprotective effects [30]. It was also shown to block the expression of several adhesion molecules, such as ICAM-1 and VCAM-1, and chemokines/receptors, such as MCP-1, MIP-1α, MIP-2, KC, and CXCR2 [31], in different disease models. However, the therapeutic effect and mechanism of cordycepin treatment in multiple sclerosis have not been investigated previously. This study is the rst to document the effect and mechanism of cordycepin in an EAE model. However, the 3D structure of cordycepin is similar to adenosine and may participate in certain biochemical reactions of adenosine. Adenosine receptors (ARs) are members of the G protein-coupled receptor family and four AR subtypes have been identi ed: A1AR, A2aAR, A2bAR, and A3AR [67]. Cordycepin may induce apoptosis in cancer cells by activating adenosine receptors [68,69]. Further investigations are needed to determine whether cordycepin binds to the various subtypes of adenosine receptors to elicit the effects we observed on the EAE model.

Conclusion
We demonstrated that cordycepin treatment ameliorated EAE by blocking immune cell in ltration in the preventive model and reducing neuroin ammation in the therapeutic model. These ndings may provide insights into the development of novel therapeutic agents for the treatment of MS.  The effects of cordycepin treatment on adhesion molecule production, chemokine receptor expression and migration of BMDCs. BMDCs were cotreated with LPS (0.1 μg/mL) with/without 50, 100, 150 μg/mL cordycepin (CO50, CO100, CO150) for 18 h. The levels of the adhesion molecules (A), integrin β1, integrin α4, c-type lectin, and ICAM-1, as well as the chemokine receptors (B) CCR2 and CCR7 on the cell surface were measured using ow cytometry. Control refers to the vehicle control (ddH2O treatment). (C) In the transwell migration assay, the upper chamber contained BMDCs from pretreated groups (LPS stimulation with or without cordycepin treatment) in 100 µl of serum-free media, while the lower chamber contained 500 µl of complete media with or without CCL21 (250 ng/ml). The migrated cells were analyzed using ow cytometry (left panel) and the WST-1 assay (right panel). Data are presented as the means + SD.  The chemokine expression pro les in the brain, spinal cord, and serum of EAE mice treated with or without cordycepin. EAE mice were administered saline (EAE-Saline) and 50 mg/kg cordycepin (EAE+CO) daily beginning on day 0 after immunization. The CFA group was similarly immunized without MOG as a negative control. Brain (A), spinal cord (B) and serum (C) samples were harvested from different groups on day 14 after immunization. Chemokine production was measured using a Mouse Chemokine Array Kit.
The bar charts show the fold change in the level of the indicated chemokine compared to the corresponding positive control spots. HSP60 was used as a positive control in the brain and spinal cord.    Proposed mechanism of action of cordycepin treatment in EAE. Cordycepin treatment induces preventive effects by inhibiting the activation of DCs, decreasing the production of tra cking factors and blocking the in ltration of immune cells, and exerts therapeutic effects by reducing neuroin ammation in the CNS.