Recombinant Human Soluble Thrombomodulin Suppresses Arteritis in a Mouse Model of Kawasaki Disease

Introduction and Objective: Kawasaki disease (KD) is associated with diffuse and systemic vasculitis of unknown aetiology and primarily affects infants and children. Intravenous immunoglobulin (IVIG) treatment reduces the risk of developing coronary aneurysms, but some children have IVIG-resistant KD, which increases their risk of developing coronary artery injury. Here, we investigated the effect of recombinant human soluble thrombomodulin (rTM), which has anticoagulant, anti-inflammatory, and cytoprotective properties on the development of coronary arteritis in a mouse model of vasculitis. Methods: An animal model of KD-like vasculitis was created by injecting mice with Candida albicans water-soluble fraction (CAWS). This model was used to investigate the mRNA expression of interleukin (IL)-10, tumour necrosis factor alpha (TNF-α), and tissue factor (TF), in addition to histopathology of heart tissues. Results: rTM treatment significantly reduces cardiac vascular endothelium hypertrophy by 34 days after CAWS treatment. In addition, mRNA expression analysis revealed that rTM administration increased cardiac IL-10 expression until day 27, whereas expression of TNF-α was unaffected. Moreover, in the spleen, rTM treatment restores IL-10 and TF expression to normal levels. Conclusion: These findings suggest that rTM suppresses CAWS-induced vasculitis by upregulating IL-10. Therefore, rTM may be an effective treatment for KD.


Introduction
Kawasaki disease (KD), an acute febrile multisystem vasculitis, is the most common cause of acquired cardiac disease in children [1]. In fact, the incidence of KD continues to increase despite the decline in the global child population. In addition, the number of KD types that do not fully conform to the diagnostic criteria is also increasing. When untreated, coronary aneurysms occur in approximately 25% of patients and have been reported to contribute to the development of cardiovascular disease in adults [2][3][4]. Although the aetiology of KD remains unknown, high-dose intravenous immunoglobulin therapy (IVIG therapy) is the first-line treatment strategy, reducing the incidence of coronary aneurysm from 25% to 1%-5% [2][3][4]. However, KD refractory to IVIG therapy is reported in approximately 10%-20% of patients and cardiovascular sequelae are reported in 2%-3% of patients. Most paediatricians prescribe additional courses of IVIG, steroids, or infliximab, and anti-tumour necrosis factor alpha (TNF-α) drugs including etanercept [2][3][4]. However, the effectiveness of these agents in controlling coronary arteritis has not yet been established [5]. Recently, the anti-inflammatory cytokine, interleukin (IL)-10, was reported to prevent vascular inflammation in a mouse model of KD [6], suggesting that anti-inflammatory cytokines can be potent agents for controlling coronary arteritis in KD.
Although no aetiological factors related to KD have been identified, epidemiological evidence suggests that environmental factors may contribute to KD susceptibility and severity, taking into consideration the seasonal variations in KD development [7]. Recently, Rodó et al. [8] identified Candida as a major fungus in tropospheric dust in north-eastern China, and found it to be associated with a KD outbreak in Japan. To date, 3 mouse models have been widely used to study the aetiology of vasculitis in KD: administration of a Candida albicans water-soluble fraction (CAWS), Lactobacillus casei cell wall extract, and a synthetic Nod1 ligand [9][10][11]. While the vasculitis observed in these murine models is similar to KD, Rodó et al. [8] suggested that CAWS-induced vasculitis is most suitable for studying the aetiology of KD.
Thrombomodulin (TM) (CD141) is an integral membrane protein expressed on the surface of endothelial cells, mesothelial cells, monocytes, and dendritic cell subsets, and serves as a cofactor for thrombin [12]. TM regulates blood coagulation by converting thrombin from a pro-coagulant enzyme to an anticoagulant enzyme. The thrombin-TM complex activates protein C to produce activated protein C, which inactivates factors VIIIa and Va in the presence of protein S, thereby inhibiting further thrombin generation [13,14]. Recombinant human soluble thrombomodulin (rTM) comprises the active extracellular domain of TM and inhibits blood coagulation by binding to thrombin [15,16]. In addition, rTM has been used in Japan since 2008 for the treatment of disseminated intravascular coagulation (DIC), a severe thrombotic disease [17]. Interestingly, rTM has been reported to have anti-inflammatory and cytoprotective functions through both activated protein C-dependent and -independent mechanisms [18,19].
In this study, we used a mouse model of CAWS-induced vasculitis to investigate whether rTM can suppress the development of coronary arteritis. This is the first report to show potency of rTM against KD therapy.

Ethics Statement
All experiments were approved by the Use and Care of Experimental Animals Committee of the Suzuka University of Medical Science Guide for Laboratory Animals Use (Licence No. 30) and were performed according to the guidelines of the Suzuka University of Medical Science. DBA/2 mice (male, 5 weeks old) were purchased from Japan SLC, Inc. (Tokyo, Japan). The mice were bred in an environment maintained at 23°C ± 1°C and 50% ± 10% relative humidity (3-6/cage); they had free access to food and water. A light/dark cycle was maintained in which the lights were turned on from 8:00 to 20:00 h.
Preparation of CAWS CAWS was prepared from C. albicans strain NBRC1385, which was acquired from the National Institute of Technology and Evaluation Biological Resource Center [11,20]. Briefly, 5 L of completely synthetic medium (C-limiting medium) containing C. albicans was cultured in a glass incubator at 27°C for 2 days, and air was supplied at a rate of 5 L/min. The culture was centrifuged at 400 rpm to remove C. albicans, and an equal amount of ethanol was added to the supernatant. The mixture was allowed to stand overnight, and then the precipitate was collected. The precipitate was dissolved in 250 mL of distilled water, and ethanol was added. The mixture was then left overnight. The resulting precipitate was collected and dried with acetone to obtain CAWS.

Effect of rTM on CAWS-Induced Vasculitis Model of Mice
The rTM, prepared as previously reported [16], was supplied by Asahi-Kasei Pharma Co., Ltd. (Tokyo, Japan). CAWS-induced vasculitis was generated as previously described [11,20]. Briefly, as shown in Figure 1, DBA/2 mice were intraperitoneally injected CAWS (1 mg/mouse/day) for 5 consecutive days until day 0 of the experiment. In the absence of pharmacokinetics data for rTM in mice, we designed the regimen of rTM administration based on other studies in which mouse models were administered ip injections [21][22][23][24], as follows: rTM was administered in 2 treatment intervals (Set A: 10 consecutive days from day 1; Set B: 5 consecutive days from days 1 and 15) and 4 doses (0, 0.2, 1, and 4 mg/kg). For comparison, the mice were injected with 0.2 mL of phosphatebuffered saline for 2 weeks after CAWS treatment (CAWS-treated group) (Fig. 1a). The mice were weighed and sacrificed by etherification on the indicated day ( Fig. 1). Autopsy was performed to obtain the heart and spleen, and both the organs were weighed without removing the blood. Half of the tissue was fixed in 10% neutralized formalin and the remaining half was stored in RNA later solution (QIAGEN N.V., Limburg, The Netherlands).

Real-Time RT-PCR Analysis
The primer pairs used in this study are shown in Table 1. Total RNA was prepared from the heart and spleen using ISOGEN (Nippon Gene Co., Ltd., Toyama, Japan), according to the manufacturer's instructions. Real-time reverse transcription (RT)-PCR analysis was performed using StepOnePlus TM (Thermo Fisher Scientific, Inc., Waltham, MA, USA) to detect mRNA expression. The expression levels of the tested genes were normalized by subtracting the corresponding threshold cycle value for glyceraldehyde-3-phosphate dehydrogenase. Normalization was performed using the ΔΔCT comparison method. DOI: 10.1159/000520717

Histopathological Evaluations
The heart was immersed in 4% buffered paraformaldehyde overnight and replaced with 70% ethanol. The specimen was then placed on a tissue processor (ASP200; Leica Biosystems, Nußloch, Baden-Württemberg, Germany). The fixed specimen was replaced with the paraffin-impregnated specimen and then embedded in a block of molten paraffin. The block was sectioned using a microtome (RM2125; Leica). The sections showing both the coronary arteries and the most severe inflammatory cell invasion were used. And the sections of each group were prepared from similar locations in the coronary arteries. The paraffin from the tissue on the slide was dissolved by Hemo-De (Leila) and ethanol treatment. The tissues were stained with haematoxylin and eosin (H&E) and elastic fibres were stained with Elastica van Gieson (EVG). The slides were sealed and observed under an optical microscope (BX51; Olympus Corporation, Tokyo Japan) at an appropriate magnification.
Quantitative evaluation of vascular inflammation was performed as previously described [25]. We divided the areas of the coronary arteries and graded the intensity of inflammation in each segment as follows: score 3, panvasculitis; score 2, inflammation involving the tunica intima and media but not spreading through the adventitia; score 1, inflammation localized to the tunica intima; and score 0, no inflammatory cell infiltration in the vascular wall.

Statistics
Data between different study groups were evaluated by Student's t test and one-way analysis of variance with Dunnett's multiple-comparison test. The GraphPad Prism 7 software (GraphPad Software Inc., La Jolla, CA, USA) was used for all statistical analyses. A p value <0.05 was considered significant.

rTM Treatment Suppresses Cardiac Hypertrophy and Arteritis in Mice Treated with CAWS
To investigate the effect of rTM on cardiac arteritis in a mouse model, the heart weight (HW) of CAWS-treated mice was measured, considering that increase in HW can link to oedema and inflammation of the heart in the CAWS-treated mice as shown in a previous report by Ohashi et al. [25]. As shown in Figure 1a, each dose of rTM was administered 5 times in a row on the designated day after CAWS treatment, and the HW of each mouse on day 34 was determined because at around day 34, we could clearly observe histological changes in the heart [11,25]. There was no significant difference in body weight (BW) among the groups (the average BW in each group was 24.1-26.5 g; data not shown). All mice survived the duration of the study. As shown in Figure 2, the HW/BW ratio of the CAWS-treated mice significantly increased (p < 0.01) compared with that of the untreated mice, as reported previously [25]. The increased HW/BW ratio of the CAWS-treated mice was reduced by treatment with rTM at both the dosage intervals, demonstrating that rTM treatment suppressed the increase in HW, induced by CAWS treatment. Dosage intervals in Set A tended to suppress the increase in HW. The dosage intervals in Set B tended to suppress the increase in HW at all doses compared with that in Set A, and a significant suppression of increase in HW was observed at the highest dose of rTM (4 mg/kg), which was the same level as that in untreated mice. The suppression of the HW/BW ratio by the highest dose of rTM in Set B was similar to that shown in the previous study, in which the HW/BW ratio was reduced by IVIG (0.59 ± 0.07 at day 28) or etanercept (0.68 ± 0.07 at day 28) treatment in CAWS-treated mice [25].
Next, the effect of rTM on the development of coronary vasculitis was histologically investigated using the highest dose of rTM (4 mg/kg). Both EVG staining showed that the elastic fibre of CAWS-treated mice was disturbed (shown in Fig. 3, middle panel), compared with that of the untreated mice (shown in Fig. 3, left panel). However, in CAWS-treated mice with rTM, the entire vascular structure was relatively conserved, whereas hypertrophic vascular endothelium was developed (shown in Fig. 3, right panel). Histological observations based on H&E staining revealed severe inflammation in samples from CAWS-treated mice, with or without rTM administration (score 3: panvasculitis; shown in Fig. 3, middle and right panels), whereas it was not observed in samples from CAWS non-treated mice without rTM (shown in Fig. 3, left panel). These results suggest that rTM treatment could suppress the disruption of the vascular structure although it could not suppress the inflammation induced by CAWS.

Time Course of Inhibitory Effect of rTM Treatment on Increased Heart and Spleen Weights in Mice Treated with CAWS
As shown in Figure 2, the dosage interval of rTM treatment in Set B suppressed the increase in the HW/BW ratio more effectively than in Set A. Therefore, we adopted Set B as the regimen in subsequent experiments. In addition to weighing the mouse heart, spleen weight (SW) was also measured as an indicator of suppressed inflammation throughout the mouse body at 20, 27, 34, and 48 days after CAWS treatment, with or without rTM, because spleen is the organ that accumulates and metabolizes im-  munocompetent cells, especially leukocyte-derived cells (Fig. 1b). The HW/BW ratio in CAWS-treated mice administered rTM remained the same as that in untreated mice, although a significant difference was not observed due to the high variation in the HW/BW ratio in CAWStreated mice on day 27 (shown in Fig. 4a). Further observation showed that suppression of the increased HW/BW ratio by rTM was lost on day 48. In contrast, the SW/BW ratio in CAWS-treated mice administered rTM on day 20 was lower than that in CAWS-treated mice, however, it was higher than that in untreated mice. After day 27, the SW/BW ratio in mice treated with rTM increased to almost the same level as that in CAWS-treated mice (shown in Fig. 4b). At later time points (days 34 and 48), rTM exerted no effect on the increased HW/BW ratio (shown in online suppl. Fig. 1; for all online suppl. material, see www.karger.com/doi/10.1159/000520717). Collectively, these results suggested that the anti-inflammatory effect of rTM was terminated, presumably, due to the rTM halflife depending on its distribution in each tissue.

Histological Changes Associated with the Time of Cardioarteritis after rTM Treatment
Histological examinations were performed to investigate the inhibitory effect of rTM on the development of vasculitis at the base of the aorta on day 20, which is the time when histological changes are generally observed [11]. Development of hypertrophic vascular endothelium, which minimizes the vascular space, was observed in the vascular structure of CAWS-treated mice, independently of rTM administration, although destruction of elastic fibres was not observed in CAWS-treated mice (EVG staining). However, the size of the vascular structure in mice without rTM administered was larger than that in mice administered rTM (shown in Fig. 5, middle and right panels). No significant differences were detected in the composition of the inflammatory cell infiltrate, which was assigned score 3 in the scoring for the extent of vasculitis and thrombosis of coronary arteries, with or without rTM administration. No changes were observed in untreated mice (shown in Fig. 5, upper panels). These results suggested that rTM does not affect the induction

Effect of rTM on the Expression of Genes Associated with Inflammation Induced by CAWS
To investigate the mechanism underlying the alleviation of coronary vasculitis by rTM, the expression of genes associated with inflammation was examined by RT-qPCR, using mRNA extracted from the heart. On day 20, Il10 expression remained unchanged in the CAWStreated mice compared with that in the untreated mice; however, the expression of Il10 in CAWS-treated mice administered rTM was significantly upregulated (p < 0.05) compared with that in untreated mice, and more significantly (p < 0.0005) compared with that in the CAWS-treated mice (shown in Fig. 6a). This increase in Il10 expression in CAWS-treated mice administered rTM was observed until day 27 (shown in Fig. 6b). Next, the gene expression levels of the pro-inflammatory cytokine, TNF-α, was examined because TNF-α is enhanced in CAWS-treated mice and may be reduced by the administration of IVIG or etanercept [25]. Although no significant difference was observed in Tnfa expression among the 3 groups on day 20 (shown in Fig. 7a), on day 27 its expression was upregulated in samples from CAWStreated mice not subjected to rTM treatment (shown in Fig. 7b). These results were consistent with histological evaluation of the heart (Fig. 3, 5), which showed panvasculitis (score 3).
As shown in Figure 4b, rTM treatment suppressed the increase in the SW/BW ratio on day 20, but not on day 27. In addition, rTM treatment significantly increased Il10 expression in the heart of CAWS-treated mice on day 20. Therefore, we investigated the effect of rTM treatment on the expression of Il10 in the spleen on day 20. The expression of Il10 in the spleen of CAWS-treated mice was downregulated compared with that in untreated mice; however, its expression in CAWS-treated mice administered rTM was restored to the same level as in untreated mice (shown in Fig. 8a). We next investigated the gene expression of thrombus-promoting tissue factor (TF), which induces thrombus formation and DIC, to confirm whether rTM could relate to the suppression of spleen hypertrophy nor not. TF is believed to be a marker of inflammation of vascular endothelial cells related to thrombomodulin because it is expressed on surface of monocyte and vascular endothelial cells in the whole body upon tissue disorder. Tf may be associated with the significant (p < 0.0005) increase in the SW/BW ratio due to spleen hypertrophy in CAWS-treated mice on day 20, as shown in Figure 4b. Tf expression in the spleen of the CAWStreated mice was significantly increased (p < 0.05) compared with that in untreated mice, whereas its expression in the spleen of CAWS-treated mice was significantly (p < 0.01) decreased to the same level as in untreated mice following treatment with rTM (shown in Fig. 8b). There was no significant difference in Tf expression in heart samples with or without tTM on days 20 and 27 (data not shown), presumably due to the low expression of Tf in the heart. These results suggest that rTM could regulate the inflammatory activation of coagulation and tissue remodelling, as reported for a study performed using a human endothelial microfluidic model [26]. No significant differences in Il10 and Tf expression in the spleen was observed in samples from CAWS-treated mice administered rTM compared with the expression in samples from CAWS-treated mice not administered rTM, consistent with the SW/BW ratio on day 27 (shown in online suppl. Fig. 2).

Discussion
KD is characterized by coronary vasculitis, followed by aneurysm, which can lead to the development of coronary artery aneurysms if left untreated with IVIG therapy, the standard treatment for acute KD. However, 10%-20% of patients are resistant to IVIG therapy and are at an increased risk of coronary vasculitis. In addition, the relative roles of secondary IVIG infusion, corticosteroids, calcineurin inhibitors, and interleukin-1 antagonists remain controversial [27]. However, recent studies have reported favourable effects for rTM on experimental and clinical DIC associated with sepsis, acute promyelocytic leukaemia, and other diseases associated with vascular inflammation and blood coagulation [15,16,19]. rTM exhibits anti-inflammatory properties in addition to anticoagulant properties.
In the current study, we, therefore, investigated whether rTM has a suppressive effect on the development of KD-related coronary vasculitis using a CAWS-induced mouse model of KD. Treatment with rTM suppressed the enlargement of vascular endothelium of the heart in KD model mice and alleviated hypertrophy in the heart and spleen (shown in Fig. 2-5). The time lag in dilation between the heart and spleen indicated that the process of vascular inflammation is a slow event in this mouse model. As shown in Figures 3 and 5, histological observation at day 20 and day 27 showed that the vascular size of CAWS-treated mice administered rTM was smaller than that of CAWS-treated mice not administered rTM (day 20) and the structure of elastic fibres was preserved in the former and was disturbed in the latter due to the enlargement of vascular endothelium (day 27). We do not have enough data to determine observed enlargement of endothelium as hypertrophy or hyperplasia. These results, however, suggest that rTM reduced the enlargement of vascular endothelium, which results in the destruction of elastic fibres in the heart, in the first 2 weeks of its administration; however, it did not fully reverse the inflammation in CAWS-treated mice. As far as we tested, no sample, where the structure of elastic fibres was destructed, was observed in CAWS-treated mice with 4 mg/kg of rTM, although it is not ruled out that 4 mg/kg of rTM can completely inhibit disruption of vascular structure. Therefore, further adjustments in the rTM treatment regimen would be needed to achieve the healing of hypertrophy. The gene expression analysis further suggested that rTM induces IL-10 production while suppressing the production of TF, resulting in the suppression of hypertrophy of vascular endothelial cells and spleen hypertrophy (shown in Fig. 6, 8). The level of IL-10 was also reported to be increased in rTM-treated dendritic cells in a previous study [28]. In addition, induction of IL-10 was observed in mouse lungs following rTM treatment (severe acute respiratory distress syndrome) and in spleen tissues (haematopoietic stem cell transplantation model) [21,22]. Therefore, rTM may specifically induce IL-10 production and suppress TF production.
Certain inflammatory cytokines, particularly TNF-α, are implicated in the pathogenesis of KD [2,25]. In fact, anti-TNF-α drugs, such as the chimeric murine/human immunoglobulin G1 monoclonal antibody infliximab, have been shown to be effective in treating patients with  KD who are refractory to initial IVIG therapy. Moreover, recent murine studies have demonstrated that IL-1 also has a significant role in driving cardiac inflammation independently of TNF-α [29,30]. In fact, multiple case reports have consistently demonstrated the successful use of anakinra, a modified human IL-1 receptor antagonist protein, in treating patients with IVIG-resistant KD [31][32][33][34]. Additionally, Nakamura et al. [6] reported a significant reduction in aortic root and coronary vascular inflammation and fibrosis in CAWS-treated mice that had been injected with adeno-associated virus-mediated IL-10 with suppressing inflammatory cytokines and development of hypertrophic vascular endothelium. We demonstrate that rTM treatment-induced IL-10 production (shown in Fig. 5, 7) whereas suppression of inflammatory cytokine including TNF-α was not observed (shown in Fig. 6). The difference might occur due to the lower concentration of IL-10 in our study.
Because we observed that rTM treatment-induced IL-10 production without altering the production of TNF-α or induction of cardiac arteritis, we would suggest that combinatorial therapy including rTM and anti-TNF-α drugs may be successful at controlling KD coronary arteritis; however, the following studies are required to confirm the efficacy of rTM in the treatment of KD patients: (1) elucidate the underlying mechanism of KD-related vasculitis and clarify the relationship between rTM and the IL-1 pathway; and (2) confirm the efficacy of rTM in other KD mouse models including the Lactobacillus casei cell wall extract model, as no animal model perfectly recapitulates KD due to our current incomplete understanding of the aetiological factors and the molecular mechanisms related to KD [35].

Conclusion
The findings of this study suggest that rTM induces IL-10 production, reduces TF production, and prevents destruction of the vascular structure in a mouse model of KD, presumably by preventing vascular endothelium enlargement. Epidemiological studies have reported that healthy individuals who maintain high plasma thrombomodulin levels have a low risk of developing coronary  artery syndrome [36,37]. In addition, rTM neutralises histones, the major extracellular mediator of death, and reduces fatal thrombosis in mice [38]. This is in line with our findings. Therefore, rTM induces IL-10 production and prevents histone-induced thrombosis, which may be an alternative therapeutic strategy for KD. In summary, our results may provide new insights into the underlying the mechanisms of KD pathogenesis.