Recombinant thrombomodulin ameliorates experimental autoimmune encephalomyelitis by suppressing high mobility group box 1 and inflammatory cytokines

Summary

Recombinant thrombomodulin (rTM) has pleiotrophic properties, including anti‐coagulation and anti‐inflammation; however, its effectiveness as a treatment for multiple sclerosis (MS) has not been evaluated fully. High mobility group box 1 (HMGB1) and proinflammatory cytokines, working as inflammatory mediators, are reportedly involved in the inflammatory pathogenesis of MS. The aim of this study was to determine whether rTM can be a potential therapeutic agent for experimental autoimmune encephalomyelitis (EAE). EAE mice received rTM treatment (1 mg or 0·1 mg/kg/day) from days 11 to 15 after immunization. The clinical variables, plasma levels of inflammatory cytokines and HMGB1 and pathological findings in EAE were evaluated. rTM administration ameliorated the clinical and pathological severity of EAE. An immunohistochemical study of the spinal cord showed weaker cytoplasmic HMGB1 staining in the rTM‐treated EAE mice than in the untreated EAE mice. Plasma levels of inflammatory cytokines and HMGB1 were suppressed by rTM treatment. In conclusion, rTM down‐regulated inflammatory mediators in the peripheral circulation and prevented HMGB1 release from nuclei in the central nervous system, suppressing EAE‐related inflammation. rTM could have a novel therapeutic potential for patients with MS.

Introduction

Multiple sclerosis (MS) is a demyelinating disease of the central nervous system (CNS) 1. A growing number of studies have revealed the important roles of inflammation, such as up‐regulation of cytokines or inflammatory mediators, in the pathogenesis of MS 2, 3, 4. Among the various cytokines, interleukin (IL)‐17 relates to the development of MS and experimental autoimmune encephalomyelitis (EAE), an animal model of MS 5, 6, 7, 8, 9. High mobility group box 1 (HMGB1) is a damage‐ or pathogen‐associated molecular pattern that works as an inflammatory mediator when released outside the cell 10, 11. We have previously reported increases of HMGB1 in serum and cerebrospinal fluid (CSF) and a positive correlation between CSF HMGB1 levels and CSF cell counts in patients with MS 4. In patients with MS, HMGB1 may help to control autoimmune responses by stimulating the release of inflammatory cytokines. In fact, administration of an anti‐HMGB1 monoclonal antibody ameliorated the clinical severity, CNS inflammation, demyelination and serum IL‐17 up‐regulation in EAE 6. Although the anti‐HMGB1 antibody has potential as a treatment for MS, it is not yet available in clinical practice.

Thrombomodulin (TM) is a membrane protein expressed on the endothelial cell surface and functions as a co‐factor in the thrombin‐induced activation of protein C in the anti‐coagulant pathway. TM reportedly has anti‐inflammatory properties: TM binds to HMGB1 and promotes the proteolytic cleavage of HMGB1 by thrombin 12. The recombinant lectin‐like domain of TM suppressed polymorphonuclear leucocyte adhesion to vascular endothelial cells and diminished cytokine‐induced increases in nuclear factor‐kappa B and activation of extracellular signal‐regulated kinase 13.

From the accumulated data, we believe that TM‐induced suppression of HMGB1 and cytokines could be a novel treatment for MS. Recombinant TM (rTM), which is composed of the extracellular portion of TM [lectin‐like domain (D1), consisting of six epidermal growth factor (EGF)‐like repeats domain (D2) and an O‐glycosylation–rich domain (D3)], is already available as a treatment for disseminated intravascular coagulation in clinical practice (Recomodulin^®; Asahi Kasei Pharma Co., Tokyo, Japan), but its efficacy as a treatment for autoimmune inflammatory diseases has not been studied. In this study, we focused on the effects of rTM on EAE to determine if rTM could be a novel therapeutic agent for MS.

Materials and methods

EAE induction

Wild‐type C57BL/6 mice (10 weeks old, female) were purchased from Japan SLC, Inc. (Shizuoka, Japan). A maximum of four animals were housed per cage and all mice had free access to water and standard rodent chow in dedicated, pathogen‐free facilities at Chiba University. A total of 200 μg myelin oligodendrocyte glycoprotein (MOG) peptides 35–55 in complete Freund's adjuvant containing killed Mycobacterium tuberculosis H37Ra was administered subcutaneously to the mice (day 1). Then, the mice received intraperitoneal injections of 400 ng of pertussis toxin (days 1 and 2). These procedures were performed using Hooke Kits (EK‐2110; Hooke Laboratories, Inc., Lawrence, MA, USA), according to the manufacturer's instructions.

All experimental animal procedures were approved by the Institutional Animal Care and Use Committee of Chiba University (approved no. 29–41).

rTM administration to EAE mice

rTM (Recomodulin^®; Asahi Kasei Pharma Co.) was prepared in sterile phosphate‐buffered saline (PBS). A total of 100 μl of rTM (1 mg/kg) (rTM^high group; n = 8), rTM (0·1 mg/kg) (rTM^low group; n = 8) or PBS alone (PBS group; n = 8) was administered intraperitoneally to EAE mice on days 11–15 after immunization with MOG, denoting the beginning of the acute phase of EAE.

Evaluations of EAE mice severity

From days 1 to 30, EAE clinical scores, body weight and rotarod time were checked every day (detailed methods are described below). These clinical parameters were evaluated by researchers blinded to the treatment status of each animal. EAE mice were scored on the following scale: 0·0 = no clinical signs; 1·0 = partial paralysis of tail (tip and base); 2·0 = limp tail and mild bilateral hind leg paralysis; 3·0 = limp tail and complete paralysis of the hind legs; 4·0 = limp tail and complete hind leg and partial front leg paralysis; and 5·0 = complete hind and front leg paralysis (rolling or dead). A mouse was assigned an ‘in‐between’ score (i.e. 0·5, 1·5, 2·5, 3·5, 4·5) when its clinical status lay between two defined scores. To assess motor function objectively, the rotarod test in EAE mice using an accelerating rotarod apparatus (Panlab Harvard Apparatus, LE8205, Barcelona, Spain) were performed. The mice were pre‐trained on the rotarod using the accelerating speed mode (the setting was from 4 to 40 rpm over 180 s), until they reached a stable baseline performance. We checked the time of the EAE mice on the rotarod twice daily, and recorded the better time. Each test trial lasted for a maximum of 80 s (the rod rotated at 20 rpm at 80 s).

Plasma cytokines and HMGB1 measurements

Blood from EAE mice was sampled on days 0 and 16. Plasma IL‐1b, IL‐6, IL‐10, IL‐17A, interferon (IFN)‐γ and tumour necrosis factor (TNF)‐α levels were measured using a multiplexed fluorescent magnetic bead‐based immunoassay (Bio‐Plex Pro™ Mouse Cytokine T helper type 17 panel A 6‐Plex; Bio‐Rad Laboratories, Hercules, CA, USA), and plasma HMGB1 levels were determined using an HMGB1 ELISA kit II (Shino‐Test Corporation, Tokyo, Japan), according to the manufacturer's instructions.

Pathological evaluations

To analyse the representative pathology in each rTM^high group and in the PBS group, two mice that had median severity scores at day 18 were killed and used for histological analyses. Untreated normal mice (n = 2) were used as controls. Pathological examinations were performed using paraffin‐embedded sections of spinal cords that were removed on day 18 after EAE induction. The spinal cord from the cervical to lumbar tissue was divided into six parts, cut along the axial plane, and the tissue was stained with haematoxylin and eosin (H&E), Luxol Fast Blue (LFB) and HMGB1. Immunohistochemical HMGB1 staining was performed using a rabbit polyclonal antibody against HMGB1 (Proteintech Group, Inc., Tokyo, Japan; species reactivity: human, mouse, rat), according to the manufacturer's instructions. Spinal cord sections were examined under a light microscope.

During the histological evaluation we examined six axial sections per mouse, from the cervical to lumbar spinal cord, and scored inflammation as follows (inflammatory index) (Fig. 1): 0 = no inflammation; 1 = cellular infiltration only in the perivascular areas and meninges; 2 = mild cellular infiltration in the parenchyma (inflammatory cells per unit area < 100 cells); 3 = moderate cellular infiltration in the parenchyma (100 cells ≤ inflammatory cells per unit area < 200 cells); and 4 = severe cellular infiltration in the parenchyma (200 cells ≤ inflammatory cells per unit area).

Figure 1.

Evaluation of inflammatory index. Inflammatory index of spinal cord in experimental autoimmune encephalomyelitis (EAE) was evaluated as follows: 0 = no inflammation; 1 = cellular infiltration only in the perivascular areas and meninges; 2 = mild cellular infiltration in the parenchyma (inflammatory cells per unit area < 100 cells); 3 = moderate cellular infiltration in the parenchyma (100 cells ≤ inflammatory cells per unit area < 200 cells); and 4 = severe cellular infiltration in the parenchyma (200 cells ≤ inflammatory cells per unit area). Black bars = 100 μm.

Statistical analyses

Statistical analyses were performed using the JMP Pro version 12.0.1 (SAS Institute Inc., Cary, NC, USA). Groups were compared using the Wilcoxon signed‐rank test for paired continuous measures and the Mann–Whitney U‐test for unpaired continuous measures, as appropriate. Spearman's rank correlation coefficient was used to test the relationship among groups. P‐values < 0·05 were considered statistically significant.

Results

rTM administration ameliorated the severity of EAE

The rTM^high and rTM^low groups exhibited significantly reduced clinical EAE scores at days 13–30 and days 19–20 (P < 0·05), respectively, compared with the PBS group (Fig. 2a). The mean ± standard error of the mean (s.e.m.) of the maximum EAE scores, cumulative EAE scores and onset days of each group (from days 1 to 30) are shown in Fig. 2b–d. The maximum EAE scores in the rTM^high (P = 0·001) and rTM^low groups (P = 0·020) were significantly lower than those in the PBS group. In the rTM^high group, the cumulative score was significantly lower (P = 0·003) and onset days were later (P = 0·006) than those in the PBS group. Similarly, the body weight of EAE mice was decreased significantly and the time of the rotarod test was significantly shorter in the PBS group compared with the rTM^high and rTM^low groups (Fig. 2e,f).

Figure 2.

Effects of recombinant thrombomodulin (rTM) treatment on experimental autoimmune encephalomyelitis (EAE) severity. (a) Clinical scores in EAE mice. EAE mice were injected intraperitoneally with 1 mg/kg of rTM (rTM^high group; n = 8), 0·1 mg/kg of rTM (rTM^low group; n = 8) or phosphate‐buffered saline (PBS) (PBS group; n = 8) on days 11–15 after immunization. Results are mean ± standard error of the mean (s.e.m.). **P < 0.01 and *P < 0.05 by the Mann–Whitney U‐test between the rTM treatment group and the PBS group. (b–d) The mean ± s.e.m. of the maximum EAE scores, cumulative EAE scores and EAE onset day of each group. The cumulative and maximum EAE scores of the rTM^high group were significantly lower than those of the PBS group. Onset day of the rTM^high group was later than that of the PBS group. **P < 0·01 and *P < 0·05 by the Mann–Whitney U‐test. (e,f) Body weight and rotarod time in each group. Results are mean ± s.e.m. **P < 0·01 and *P < 0·05 by the Mann–Whitney U‐test between the rTM treatment group and the PBS group.

rTM administration suppressed up‐regulation of inflammatory cytokines in EAE

Plasma levels of IL‐1b, IL‐6, IL‐10, IL‐17A, IFN‐γ and TNF‐α were up‐regulated significantly at day 16 compared with day 0 in all groups. However, the up‐regulation of IL‐1b, IL‐6, IL‐17A and IFN‐γ was attenuated in the rTM^high group compared with the PBS group (Fig. 3).

Figure 3.

Effects of recombinant thrombomodulin (rTM) treatment on plasma cytokine levels in experimental autoimmune encephalomyelitis (EAE). Plasma cytokine levels of EAE mice that received phosphate‐buffered saline (PBS) (PBS group; n = 8), 1 mg/kg of rTM (rTM^high group; n = 8) or 0·1 mg/kg of rTM (rTM^low group; n = 8) on days 0 (black bar) and 16 (white bar) are shown. The inflammatory cytokines were elevated on day 16 compared with day 0 in all groups (**P < 0·01 and *P < 0·05 by the Wilcoxon test). Plasma levels of interleukin (IL)‐1b, IL‐6, IL‐17A and interferon (IFN)‐γ were lower in the rTM^high group than in the PBS group (^† P < 0·05 and ^‡ P < 0·10 by the Mann–Whitney U‐test).

IL‐17A and HMGB1 were involved in developing EAE

The IL‐17A ratio (defined as the ratio of IL‐17A level on day 16 to that on day 0) was lower in the rTM^high group than in the PBS group (Fig. 4a). Only the IL‐17A ratio showed a significant positive correlation with the cumulative EAE score in the PBS group (P = 0·028, r = 0·762) (Fig. 4b) among the measured cytokines (IL‐1b, IL‐6, IL‐10, IL‐17A, IFN‐γ and TNF‐α). This suggests that IL‐17A contributes to the development of EAE in the peripheral circulation. Plasma HMGB1 levels were increased at day 16 compared with day 0 in the PBS group, whereas they were decreased in the rTM‐treated groups; however, there was no statistical significance among them (Fig. 4c). The ratio of HMGB1 (defined as the ratio of HMGB1 level on day 16 to that on day 0) was correlated significantly with that of IL‐17A in the PBS group (P = 0·047, r = 0·714) (Fig. 4d).

Figure 4.

Plasma interleukin (IL)‐17A and high mobility group box 1 (HMGB1) levels in experimental autoimmune encephalomyelitis (EAE). (a) The IL‐17A ratio was lower in the recombinant thrombomodulin (rTM)‐treated EAE group than in the phosphate‐buffered saline (PBS) group (n = 8 in each group). (b) The IL‐17A ratio (IL‐17A level on day 16/day 0) showed a significant positive correlation with the cumulative EAE score in the PBS group (n = 8). (c) Plasma HMGB1 levels on day 16 were increased in the PBS group, whereas they were decreased in the rTM‐treated EAE group (n = 8 in each group). There was no statistical significance among them. (d) The HMGB1 ratio (HMGB1 level on day 16/day 0) was associated significantly with the IL‐17A ratio in the PBS group (n = 8).

rTM administration attenuated CNS inflammation and demyelination in EAE

rTM treatment (1 mg/kg, days 11–15; rTM^high group) reduced the infiltration of cells and demyelination in the spinal cord of EAE mice and reduced the inflammatory index [rTM^high group = 1·42 ± 0·15 (mean ± s.e.m.); PBS group = 2·33 ± 0·22; P = 0·006) (Fig. 5). Immunohistochemical analyses demonstrated increased extra‐nuclear cytoplasmic HMGB1 staining in the EAE + PBS group compared with that observed in normal mice, whereas extra‐nuclear cytoplasmic HMGB1 staining was unremarkable in the EAE + rTM^high group (Fig. 5).

Figure 5.

Pathological findings of the spinal cord in experimental autoimmune encephalomyelitis (EAE). Haematoxylin and eosin (HE), Luxol Fast Blue (LFB) and high mobility group box 1 (HMGB1) staining of the spinal cord of normal mouse, EAE mouse receiving 1 mg/kg of recombinant thrombomodulin (rTM) (EAE + rTM^high group), and EAE mouse receiving phosphate‐buffered saline (PBS) (EAE + PBS group) was performed (n = 2 in each group, representative images are shown). Inflammation and demyelination in the EAE + rTM^high group were ameliorated compared with the EAE + PBS group. The inflammatory index of the EAE + rTM^high group was significantly lower than that of the EAE + PBS group. HMGB1 immunostaining showed outstanding cytoplasmic staining in the EAE + PBS group and mild cytoplasmic staining in the EAE + rTM^high group compared with normal mouse. Blue bars = 200 μm.

Discussion

This is the first study, to our knowledge, on the efficacy of rTM treatment in MS using EAE. rTM could have a dual function for treating EAE: in the peripheral blood circulation, it suppressed IL‐17 (and, to a lesser extent, HMGB1 and other inflammatory cytokines); in the CNS, it prevented HMGB1 release from the nuclei, which further suppressed inflammation expansion. rTM worked as an anti‐inflammatory agent in both the periphery and in the CNS, ameliorating the clinical signs of EAE (Fig. 6).rTM has pleiotrophic roles, including anti‐inflammation. rTM has a lectin‐like domain (D1), which binds HMGB1 14, accelerates HMGB1 fragmentation with thrombin 12, inhibits neutrophil adhesion to endothelial cells 13, 15 and suppresses complement activation 16. rTM also activates thrombin‐activatable fibrinolysis inhibitors, leading to inhibition of complements 16. Furthermore, thrombin activates protein C by binding D2 of rTM. Activated protein C has cytoprotective effects including anti‐inflammation, anti‐apoptosis and endothelial barrier protection 17. Recently, it has been reported that protein C activation was required for mitochondrial function and myelination in the CNS; impaired protein C activation aggravated EAE, which can be compensated for by soluble TM, partially independent of immunomodulation 18. In this study, rTM suppressed the up‐regulation of plasma HMGB1 and inflammatory cytokines (including IL‐17), which could ameliorate clinical symptoms of EAE. HMGB1, known as an inflammatory inducer and secreted by immune cells (pathogen‐associated molecular pattern) or damaged cells (damage‐associated molecular pattern), is important for EAE and MS pathogenesis, as there is a potential interaction between HMGB1 and CNS inflammation 4, 6. Consequently, HMGB1 blockade by rTM could be a therapeutic target for MS. Among inflammatory cytokines, IL‐17 played a crucial and predominant role in the development of EAE 5. We have reported previously that IL‐17 levels were up‐regulated significantly in EAE and treatment with an anti‐HMGB1 monoclonal antibody reduced IL‐17 production in the peripheral circulation, thereby lessening the clinical and pathological severity of EAE 6. In this study, the IL‐17A and HMGB1 ratios were correlated significantly in EAE mice. From these findings, HMGB1 inhibition could contribute directly to IL‐17 suppression. Further studies are needed to confirm the definite mechanism between HMGB1 and IL‐17.

Figure 6.

Effects of recombinant thrombomodulin (rTM) treatment on experimental autoimmune encephalomyelitis (EAE) pathogenesis. In the periphery, interleukin (IL)‐17, high mobility group box 1 (HMGB1) and inflammatory cytokines were associated with EAE aggravation. HMGB1 and IL‐17 were associated with each other. rTM mainly suppresses IL‐17, leading to EAE attenuation. In the central nervous system (CNS), HMGB1 from damaged or immune cells is involved in the inflammation of EAE. rTM suppresses HMGB1, resulting in amelioration of EAE.

In the CNS of EAE, rTM administration decreased the invasion of inflammatory cells, demyelination and HMGB1 release to cytoplasm. As we have reported previously, CNS inflammation induced HMGB1 release from the cell nuclei, promoting further CNS inflammation in EAE 6. In this study, rTM inhibited HMGB1 release from the nuclei into the cytoplasm of the cells around the inflammatory lesion, suppressing inflammation and demyelination in the CNS of EAE mice, similar to anti‐HMGB1 antibody treatment 6. To reveal a more definite role of rTM in the CNS of EAE, the analyses of mRNA expression of cytokines in the spinal cord or cytokine levels in the cerebrospinal fluid will be required.

Taken together, rTM appears to have therapeutic potential for MS. Considering the molecular weight of rTM, rTM therapy may be potentially useful during acute disease exacerbations when the blood–brain barrier is disrupted.

Some limitations in our study need to be addressed. (1) rTM suppressed EAE clinical score in a dose‐dependent manner; whereas inhibitory effect on peripheral productions of cytokines and HMGB1 were not. Possible explanations are that effects of rTM in periphery are weaker than that in CNS, suppression of cytokines and HMGB1 occurs even at low doses of rTM and rTM has other actions in EAE besides anti‐inflammation. (2) The change of plasma HMGB1 levels showed the opposite trend in the PBS group and rTM groups (plasma HMGB1 levels on day 16 were increased compared to day 0 in the PBS group; conversely, they were decreased in the rTM groups); however, HMGB1 levels did not change significantly by rTM treatment. This may indicate that rTM primarily suppresses plasma IL‐17, not HMGB1, in the pathogenesis of EAE. (3) Although no apparent haemorrhagic side effects were observed in this study, there is a possibility of bleeding triggered by rTM treatment, because combined rTM/thrombin activates protein C, producing anti‐coagulant effects. Further studies are needed to assess this risk.

In conclusion, rTM administration to EAE showed pleiotrophic effects and ameliorated the clinical disease severity significantly. rTM treatment could be a new treatment for patients with MS and other autoimmune‐mediated inflammatory disorders.

Disclosure

The authors declare that there are no conflicts of interest.

Author contributions

All authors were involved in drafting the article or revising it critically for important intellectual content and have read and approved the final version of the manuscript.