Leonurine suppresses neuroinflammation through promoting oligodendrocyte maturation

Abstract Focal inflammation and remyelination failure are major hallmarks of multiple sclerosis and its animal model, experimental autoimmune encephalomyelitis (EAE). In this study, we found that leonurine, a bioactive alkaloid, alleviated EAE disease severity along with reduced central nervous system inflammation and myelin damage. During the pathogenesis of EAE, leonurine dramatically suppressed the recruitment of encephalitogenic T cells into the central nervous system, whereas did not impair periphery immune responses and microglia activation. Mechanistically, leonurine protected mice against demyelination along with enhanced remyelination through promoting the maturation of oligodendrocytes in both EAE and cuprizone‐induced demyelination mouse models. Moreover, we identified that the expression of demethylase jumonji domain‐containing protein D3 was significantly enhanced upon treatment of leonurine, which suppressed the trimethylation of histone H3 lysine‐27 and enhanced oligodendrocyte maturation accordingly. Collectively, our study identified the therapeutic effect of leonurine on EAE model, which potentially represents a promising therapeutic strategy for multiple sclerosis, even other demyelination disorders.

destruction of oligodendrocytes (OLs) and neurons, and then contributing to the damage of the myelin and axons. 3,4 Pharmacological strategies to modulate CNS inflammation can enable myelin repair and prevent subsequent neurodegeneration in this disease. Pathological studies also demonstrated that the endogenous myelination capacity decreases in MS patients. [5][6][7] Boosting OL differentiationmediated remyelination to replenish damaged myelin is gaining more attention in MS therapeutics discovery. 8,9 In healthy brains, myelin is continually generated and dynamically remodelled by OLs, which are differentiated from oligodendrocyte progenitor cells (OPCs). [10][11][12] During remyelination, OPCs migrate to the demyelinated lesion site, and then differentiates into OL and regenerates demyelinated myelin, protecting demyelinated axons from immune attack and irreversible degeneration. 13 However, no drugs targeting remyelination has been reported until now. 9 Exploring strategies to protect OLs and myelin during neuroinflammation is critical for MS therapy.
Leonurine (4-guanidino-n-butyl syringate), a kind of bioactive alkaloid extracted from Herba leonuri, 14 has been shown to exhibit anti-inflammatory function. For example, Song et al showed that leonurine inhibited cyclooxygenase-2 expression in lipopolysaccharideinduced mouse mastitis. 15 Li et al demonstrated that leonurine attenuated fibroblast-like synoviocyte-mediated synovial inflammation and joint destruction in rheumatoid arthritis. 16 More recently, leonurine reportedly could decrease microglia or macrophage overactivation in Alzheimer's disease and monosodium urate crystal-induced inflammation. 17,18 Besides, leonurine was surprisingly found exhibiting neuroprotective function in neurodegenerative diseases, such as Parkinson's disease, 19 stroke, 20,21 and Alzheimer's disease, 17 protecting brain from ischaemic injury and reducing neuron loss. In view of these findings, we set out to examine a possible role of leonurine in the CNS autoimmune disease EAE.
In this study, we firstly explored the therapeutic effect of leonurine on EAE mice, and found that leonurine alleviated disease severity of EAE reducing CNS inflammation and myelin damage. Leonurine reduced recruitment of encephalitogenic T cells into the CNS through suppressing CNS chemokine expression, whereas did not impair peripheral immune responses and microglia/macrophage activation. Interestingly, leonurine could protect mice from demyelination and enhance remyelination process, which was further validated in cuprizone-induced demyelination mice and may confer leonurine-ameliorated CNS inflammation in EAE model. This effect of leonurine on myelination was exerted by promoting OL differentiation partially through histone H3 lysine-27 (H3K27) demethylase jumonji domain-containing protein D3 (JMJD3). Therefore, our study demonstrates leonurine as a potential therapeutic compound for CNS autoimmune diseases.

| Induction and treatment of EAE model
Eight-week-old C57BL/6 male mice were purchased from the Shanghai Laboratory Animal Center of the Chinese Academy of Sciences, and kept under specific pathogen-free conditions in the animal centre of Institute of Health Sciences, Chinese Academy of Sciences. Agematched mice were subcutaneously immunized with peptide MOG  (GL Biochem, Shanghai, China) mixed 1:1 with complete Freund's adjuvant (Sigma, St. Louis, MO, USA), followed by pertussis toxin (200 ng; List Biological Laboratories, Campbell, CA, USA) administration i.v. on day 0 and day 2, as previously described. 22 Leonurine is chemically synthesized (Shifeng Biological Technology Company, Shanghai, China), and the purity is above 98%. For therapeutic protocol of EAE mice, leonurine (60 mg/kg) or vehicle (dimethyl sulphoxide) was daily administered (i.p.) from day 11 post immunization. Adoptive transfer EAE model was established as previously described. 23,24 Briefly, draining lymph node (DLN; proper axillary and accessory axillary) cells 25 and splenocytes isolated from vehicle-or leonurine-treated EAE mice on day 15 post immunization were re-challenged with the MOG 35-55 (10 mg/mL) peptide for 3 days. 5 × 10 6 viable cells were then transferred into sublethally irradiated mice (6 Gy

| Histological analysis
Vehicle-and leonurine-treated EAE mice were killed on day 28 post immunization. After cardiac perfusion with 4% paraformaldehyde, spinal cords were isolated, dehydrated with alcohol and embedded in paraffin. Spinal cord sections were then stained with haematoxylin and eosin (H&E) or luxol fast blue (LFB). Inflammation and demyelination scores were valued as previously described. 26

| Cell population analysis by flow cytometry
Spinal cord, brain, and DLN were isolated from vehicle-and leonur-

| Primary microglia and cell line culture
Primary microglia cultures were prepared as previously described. 23,24 Briefly, cerebral cortical cells from newborn C57BL/6 mice were iso-

| Immunofluorescence
After cardiac perfusion with 4% paraformaldehyde, spinal cords or corpus callosum of mice were isolated, dehydrated with alcohol and
After 3 weeks treatment, the brain sections were prepared, and stained with LFB or immunofluorescence (IF) stained with NG2 and CC1. Quantitative analysis of myelination by LFB staining of corpus callosum in the brain of cuprizone-induced demyelinated mice was carried out using Image-Pro Plus (NIH, Bethesda, MD, USA) as T A B L E 1 Specific primers used in real-time PCR analysis

Gene
Primer

| Murine OPC preparation and OL differentiation
Oligodendrocyte progenitor cells used in this study were differentiated from murine cortical neural stem cells (NSCs), which were isolated as described previously 31 with modification. Briefly, murine cortical NSCs were prepared from the brain on embryonic day 14.

| Ex vivo cerebellar slice culture
Cerebellar slices were isolated and cultured according to the protocol previously described. 32 Briefly, whole cerebellum was collected at postnatal day (PD) 7 and cut into 300 μm sagittal slices on a microtome (Leica, Wetzlar, Germany). Slices were cultured in a DMEM/F12 supplemented with 15% heat-inactivated horse serum (Gibco), B27 without vitamin A and platelet-derived growth factor-AA. After 1 day, the culture medium changed into DMEM/F12 with 15% heat-inactivated horse serum and B27 without vitamin A. Cerebellar slices were then treated with leonurine (5 μmol/L) or vehicle for 1 or 2 days. The culture medium was changed daily and slices were lysed for immunoblotting.

| Lentiviral vector construction
Oligonucleotides with the listed nucleotide sequences in Table 2 were used for the cloning of shRNA-encoding sequences into a len-

| Statistical analysis
All measurement data are presented as mean ± SEM. SPSS software, version 20 (IBM, Armonk, NY, USA) was used for all statistical analyses. Significant differences were evaluated using an independentsamples t test or Wilcoxon rank test. One-way ANOVA was used to determine multigroup differences, and then we adopted the Bonferroni's or Dunn's Multiple Comparison Test for further analysis to determine the difference between two groups. Significance was expressed as: *P < 0.05, and **P < 0.01.

| Leonurine ameliorates clinical severity of EAE mice
To evaluate the potential effect of leonurine on MS, we tested the therapeutic effect of leonurine on MOG 35-55 -induced EAE mice. The molecular structure of leonurine was shown in Figure 1A. We treated the EAE mice with leonurine through injection (i.p.) from day 11 post immunization, which represented disease onset in EAE mice.
The results revealed that leonurine treatment substantially alleviated the disease severity of EAE mice compared with vehicle treatment ( Figure 1B). The observed therapeutic effects of leonurine were consistent with much less infiltration of inflammatory cells and fewer  Figure 1C,D). Besides, the absolute numbers of MNCs in the CNS were also decreased after leonurine treatment ( Figure 1E). These data suggested that leonurine alleviated disease severity of EAE, reducing CNS inflammation and myelin damage.

| Leonurine prevents the recruitment of encephalitogenic T cells into the CNS
The infiltrated autoreactive T cells are known required for the induction of CNS inflammation and pathology of EAE. 4,33 Consistent with the histological results, we found that the percentages and absolute numbers of CD4 + and CD8 + T cells in the CNS of EAE mice were dramatically decreased after leonurine treatment (Figure 2A). Accordingly, the infiltration of the pathogenic IFN-γ-and IL-17-producing CD4 + T cells were also significantly decreased in the CNS of EAE mice after leonurine treatment ( Figure 2B).
We next explored whether leonurine impaired encephalitogenic T cell response in the periphery. Unexpectedly, the percentages and absolute number of CD4 + and CD8 + T cells were similar in the DLN of vehicle-and leonurine-treated EAE mice ( Figure 3A). There were also no considerable changes of IFN-γ-and IL-17-producing CD4 + T cells and Treg cells in the DLN of EAE mice after leonurine treatment ( Figure 3B). In addition, leonurine treatment had no obvious effects on the proliferation and cytokine production of the encephalitogenic T cells that re-stimulated with MOG peptides in vitro compared with vehicle controls (Figure 3C,D). Accordingly, the recipient mice adoptively transferred with peripheral encephalitogenic T cells from vehicle-or leonurine-treated EAE mice, developed similar typical EAE clinical symptoms ( Figure 3E). These results suggested that the ameliorated EAE pathology conferred by leonurine was not caused by an impaired encephalitogenic T cell response in the periphery, but due to the inhibited infiltration of pathogenic autoimmune T cells into the CNS.

| Leonurine does not affect the inflammatory response in microglia
The Cxcl10, and Ccl20 were reduced significantly in the CNS of leonurine-treated mice compared with vehicle-treated mice ( Figure 4A,B).
Accordingly, the percentages and absolute numbers of CXCR3 + / CCR5 + CD4 + encephalitogenic T cells were significantly higher in the DLN of leonurine-treated mice than vehicle-treated mice ( Figure 4C).
Thus, leonurine seemed to function directly in CNS-resident cells to suppress local neuroinflammation, and then inhibit the recruitment of autoimmune T cells into the CNS during EAE pathogenesis.
In CNS of EAE mice, activated microglia/macrophage produce proinflammatory factors that are toxic to the CNS and chemokines that promote inflammatory infiltration into the CNS. 3,4,36 We therefore explored the effects of leonurine on proinflammatory activities of CNS-resident cells such as microglia and infiltrated macrophage.
To examine the modulatory effect of leonurine on microglia, primary cultured microglia were prepared and activated with IFN-γ.
The mRNA abundance for Tnfa, Il1b, Il6, Il12b, Il23a, Ccl2, Ccl3, Ccl5, Cxcl10, and Ccl20 were significantly upregulated in IFN-γ-treated microglia. However, this upregulation was not affected by leonurine addition to microglia cultures ( Figure 4D). Besides, leonurine also did not change the production of IFN-γ-induced proinflammatory cytokine and chemokines in the murine macrophage cell line RAW 264.7 ( Figure 4E). Astrocytes have also been reported producing chemokines and playing a dominant role in the regulation of leucocyte recruitment. [37][38][39][40] Primary astrocytes were isolated, and our results showed that IFN-γ-up-regulated mRNA expression of Ccl2, Ccl3, Ccl5, Cxcl10, and Ccl20 was also not affected by leonurine addition to astrocytes cultures ( Figure S1). These results collectively demonstrated that the suppressed EAE by leonurine was not due to the inhibition of microglia/macrophage inflammatory activation.

| Leonurine promotes remyelination in EAE mice
The amelioration of EAE pathology always results from two major aspects of function, the suppression of CNS inflammation and the promotion of remyelination, 41 which promoted us to explore whether leonurine directly affected remyelination, and then conferred leonurine-ameliorated CNS inflammation in EAE mice. We then analysed the expression of myelin proteins or genes, such as MBP and proteolipid protein (PLP), which were increased during remyelination and believed to reflect an important step leading to the remyelination process. 42 As shown in Figure 5A, IF staining of spinal cords showed that nearly 2 folds of MBP expression was Remyelination is a complicated process that OPCs differentiate into immature OLs and consequently mature OLs, which possesses the capability of myelination. 43 Thus, we utilized stage-specific markers, such as NG2 for OPCs and CC1 for newly generated OLs, to evaluate remyelination in the spinal cords of EAE mice on day 28 post immunization after leonurine treatment. Sporadic NG2 + OPCs were observed in the spinal cords of vehicle-and leonurine-treated mice, and leonurine treatment reduced the number of NG2 + OPCs in the inflammatory loci ( Figure 5C). However, there were much more CC1 + newly generated OLs in the inflammatory loci of the spinal cords of leonurine-treated mice than that of vehicle-treated mice ( Figure 5D). The mRNA expression of Ng2 and Cc1 in the CNS displayed similar pattern ( Figure 5E). These results implied that leonurine may promote the remyelination in the CNS of EAE mice through enhancing OPC differentiation into mature OLs.

| Leonurine enhances remyelination in cuprizone-induced demyelination model
We next analysed the effects of leonurine on remyelination using cuprizone-induced nonimmune-mediated demyelination model, in which cuprizone induces the cell death of OLs while spares OPCs. [27][28][29] After fed with a 0.2% (w/w) cuprizone diet for 3 weeks, mice were given standard chow for 2 more weeks. Significant demyelination was induced after 2-week standard chow, and spontaneous remyelination occurred after that. Vehicle or leonurine was administered (i.p.) when spontaneous remyelination started ( Figure 6A). Although spontaneous remyelination occurred in vehicle-treated mice, remyelination in leonurine-treated mice was significantly enhanced as compared with that of control mice ( Figure 6B)

| Leonurine promotes OL differentiation through increasing JMJD3 in vitro
To further confirm the prodifferentiation effect of leonurine on OLs, we analysed OL differentiation in vitro, in which OLs were differentiated from mouse embryonic NSC-derived OPCs as described previously. 30,31 The isolated NSCs formed neurosphere ( Figure S2A) and expressed NSC specific markers, such as Olig2, Sox10, and Nestin ( Figure S2B). OPCs and OLs were generated from NSCs and validated by Pdgfra, Mbp, and Plp expression ( Figure S2C). Although incapable of inducing OL differentiation alone, leonurine dramatically enhanced OL differentiation in the presence of T3, which is a potent OL differentiation inducer, 44,45 evidenced by elevated MBP expression ( Figure 7A,B). The mRNA levels of Cnp and Plp specific for OLs were also dramatically elevated by leonurine in the presence of T3 ( Figure 7C). However, BrdU incorporation assay showed that leonurine had no obvious effects on OL proliferation ( Figure 7D). The observation that leonurine-promoted OL differentiation raised the question that whether it also promoted further myelination. We therefore set up a postnatal myelination system by culturing cerebellar slices of postnatal mice with leonurine. Cerebellar slices were prepared from mice at PD 7, when widespread myelination is not complete, 32 and treated ex vivo with leonurine or vehicle. The results showed that leonurine significantly boosted MBP expression ( Figure 7E). Taken together, our data established leonurine as a critical positive regulator of the OL differentiation and myelination.
To dissect the molecular mechanism by which leonurine promotes OL differentiation, we investigated the effect of leonurine on the methylation alteration in histone 3, which is involved in OL differentiation. 46,47 As shown in Figure 7F, we found that leonurine treatment specifically inhibited H3K27me3, whereas had no effect on H3K9me3. The methyltransferase EZH2 and demethylase JMJD3 are well-known to regulate the diversified methylation status of H3K27. Interestingly, leonurine treatment did not affect the expression of EZH2, but dramatically enhanced the expression of JMJD3.
In addition, enzymatic inhibition of EZH2 by GSK126 during OL differentiation did not impede the enhancement effect of leonurine, whereas the JMJD3-specific inhibitor GSK-J4 strongly abrogated the enhanced effect of leonurine on OL differentiation ( Figure 7G). To further validate the role of JMJD3 in leonurine function, JMJD3 was knocked down in NSCs using a lentivirus-expressing shRNA specific to Jmjd3 (named shJmjd3-1 and shJmjd3-2), and NSCs infected with a lentivirus-expressing scrambled shRNA (named shNC) were used as controls ( Figure 7H). As expected, JMJD3 knockdown also dampened the OL prodifferentiation effect of leonurine ( Figure 7I). Thus, our data suggested that leonurine elevated the OL differentiation dependent on increased JMJD3 expression.

| DISCUSSION
The therapy for EAE and MS always includes two major aspects: the suppression of CNS inflammation and the promotion of remyelination. 41 Our previous studies have demonstrated two therapeutic compounds, baicalein and 18β-glycyrrhetinic acid, reduced microgliamediated CNS inflammation and demyelination and then ameliorated EAE. 23,24 In this study, we provided several lines of evidence that leonurine alleviated disease severity of EAE by promoting OL differentiation, which conferred reduced encephalitogenic T cells Demyelination induced by autoreactive T cell response has long been considered critical for MS and EAE model. Leonurine reportedly displayed anti-inflammatory function. [15][16][17][18] However, we found that leonurine had no effect on T cell response in the periphery, indicating by similar components of CD4 + and CD8 + T cells in the DLN of vehicle-and leonurine-treated EAE mice, as well as similar proliferation, cytokine profile, and ability to induce EAE of MOG  challenged T cells from vehicle-and leonurine-treated EAE mice.
Nevertheless, leonurine could inhibit T cell infiltration in the CNS.
We further found that reduced T cell infiltration was associated with reduced chemokine in the CNS. These chemokines were mainly produced by CNS-resident cells such as microglia and infiltrating immune cells, which also produced toxic proinflammatory factors to CNS. Hong and Liu et al showed that leonurine could decreased microglia or macrophage overactivation in β-amyloid 1-40 (Aβ 1-40 )induced cognitive impaired rats. 17,18 However, our data showed that leonurine did not affect the expression of proinflammatory cytokines and chemokines in microglia/macrophage in vitro. Besides, many studies also demonstrated that astrocytes produce chemokines and play a dominant role in the regulation of leucocyte recruitment. [37][38][39][40] We showed that leonurine did not affect the expression of Ccl2, In summary, our findings demonstrate that leonurine promotes OL differentiation to protect mice from demyelination and enhance remyelination accompanied with reduced T cell infiltration in the CNS, which potentially provides a promising therapeutic strategy for MS, and even other demyelination disorders.

ACKNOWLEDG EMENTS
This work is supported by the National Natural Science Foundation of China (81471217, 81670540, 81471572, 31670140), Ministry of F I G U R E 7 Leonurine augments in vitro OL differentiation through augmenting JMJD3 signaling. OL differentiation of OPCs derived from NSCs was induced as described in Section 2. (A, B) OPCs cultured in OL differentiation medium were treated with vehicle or leonurine (5 μmol/L) in the presence of T3 or not for 5 days. Cells were IF stained for MBP expression. Quantification of the percentages of MBP + cells per field was shown (n = 4). Scale bars, 100 μm. MBP protein expression was detected by immunoblotting analysis, and the densitometry of the bands were quantified using ImageJ software (n = 4). (C) OPCs cultured in OL differentiation medium were treated with vehicle or leonurine (5 μmol/L) in the presence of T3 for 5 days. Expression of Cnp and Plp was measured by real-time PCR (n = 4). (D) BrdU analyses of proliferation of differentiated OLs treated with vehicle or leonurine (5 μmol/L) at the indicated times. (E) Cerebellar slices were isolated and cultured according to the protocol previously described. Cerebellar slices were then treated with leonurine (5 μmol/L) or vehicle for 1 or 2 days and slices were lysed for immunoblotting. The densitometry of the bands was quantified using ImageJ software (n = 4). (F) OPCs cultured in OL differentiation medium were treated with vehicle or leonurine (5 μmol/L) in the presence of T3 for 5 days. H3K27me3, H3K9me3, Histone3, EZH2, and JMJD3 expression were detected by immunoblotting analysis, and the densitometry of the bands were quantified using ImageJ software (n = 4). (G) OPCs cultured in OL differentiation medium were treated with vehicle or leonurine (5 μmol/L) in the presence of T3 for 5 days, combined with GSK-J4 or GSK-126. MBP expression was detected by immunoblotting analysis, and the densitometry of the bands were quantified using ImageJ software (n = 4). (H, I) JMJD3 expression in shNC, shJmjd3-1, and shJmjd3-2 NSCs was detected by immunoblotting analysis. Differentiated OLs were induced from shNC, shJmjd3-1, and shJmjd3-2 NSCs, and MBP expression was detected by immunoblotting analysis. The densitometry of the bands was quantified using ImageJ software (n = 4). Statistical significance indicated as *P < 0.05, **P < 0.01