All-trans-Retinoic Acid Alleviates Coronary Stenosis by Regulating the Migration of Smooth Muscle Cells in a Mouse Model of Kawasaki Disease

Coronary artery (CA) stenosis is a detrimental and often life-threatening sequela in Kawasaki disease (KD) patients with coronary artery aneurysm (CAA). Therapeutic strategies for these patients have not yet been established. All-trans-retinoic acid (atRA) is a modulator of smooth muscle cell functions. The purpose of this study was to investigate the effect of atRA on CA stenosis in a mouse model of KD. Lactobacillus casei cell wall extract (LCWE) was intraperitoneally injected into 5-week-old male C57BL/6J mice to induce CA stenosis. Two weeks later, the mice were orally administered atRA (30 mg/kg) 5 days per week for 14 weeks (LCWE+atRA group, n=7). Mice in the untreated group (LCWE group, n=6) received corn oil alone. Control mice were injected with phosphate-buffered saline (PBS, n=5). Treatment with atRA signicantly suppressed CA inammation (19.3±2.8 vs 4.4±2.8, p<0.0001) and reduced the incidence of CA stenosis (100% vs 18.5%, p<0.05). In addition, atRA suppressed the migration of human coronary artery smooth muscle cells (HCASMCs) induced by platelet-derived growth factor subunit B homodimer (PDGF-BB). In conclusion, atRA dramatically alleviated CA stenosis by suppressing SMC migration. Therefore, it is expected to have clinical applications preventing CA stenosis in KD patients with CAA.


Introduction
Kawasaki disease (KD) is an acute systemic vasculitis of unknown etiology that is mainly associated with coronary artery aneurysms (CAAs) and occurs primarily in young children 1,2 . A serious complication of CAA is CA stenosis, which includes thrombotic occlusion of CA in the acute phase and progressive CA narrowing due to intimal hypertrophy in the chronic phase. These pathological features of CA stenosis were actually identi ed in autopsy cases 3 . Recently, Fukazawa R et al. performed a nationwide survey of KD patients with giant CAA in Japan. MI occurred in 18% of these patients, and severe cardiac events were likely to occur within 2 years of the onset of KD 4 . More recently, Miura et al. identi ed the risk factors for cardiac events in KD patients with CAA. Male sex and intravenous immunoglobulin (IVIG)-resistance were independent risk factors for acute coronary events 5 . Although these clinical studies revealed the natural history of KD patients with CAA, novel therapeutic strategies for these patients have not yet been investigated.
Recent clinical studies have reported useful therapeutic tools for anti-in ammatory agents such as IVIG 6 , prednisolone 7 , in iximab 8 and ciclosporin 9 , especially during the acute phase. Therefore, according to the 25th Japanese nationwide Kawasaki disease survey, the incidence of cardiac sequelae was reduced to 2.6% in KD patients. 10 However, until now, due to the lack of a suitable animal model, no drug has been developed that is useful for improving the prognosis of KD patients who already have CAA. Pathological investigations of KD autopsy cases reported by Orenstein et al demonstrated that luminal myo broblastic proliferation (LMP) was associated with unique smooth muscle cell (SMC)-derived myo broblasts that also caused progressive CA stenosis 11 . Of note, we recently successfully discovered a novel mouse model of CA stenosis induced by Lactobacillus casei cell wall extract (LCWE) to mimic KD. Histologically, this model is characterized by intimal hypertrophy due to SMC proliferation and migration following severe CA vasculitis, leading to CA stenosis 12 . Since the LCWE-induced vasculitis model closely resembled the pathological features found in human KD autopsy cases, we therefore decided to use this mouse as a model for chronic CA stenosis.
All-trans-retinoic acid (atRA) is a natural derivative of vitamin A that inhibits cell proliferation and migration and has anti-in ammatory properties [13][14][15] . Currently, atRA is a conventional therapy for the management of acute promyelocytic leukemia (APL) 16 . Recent experimental studies have reported that atRA treatment signi cantly reduces the formation of atherosclerosis in a high-fat diet-induced rabbit model 17 . Moreover, it has also been reported that atRA reduces neointimal formation in the carotid artery after balloon withdrawal injury in a rat model 18 . In addition, the synthetic retinoid Am80 signi cantly ameliorates Candida albicans water-soluble fraction (CAWS)-induced vasculitis through the inhibition of neutrophil migration and activation 19 . Based on previous animal studies, we hypothesized that atRA could suppress CA stenosis by modulating the properties of vascular smooth muscle cells (VSMCs) in a KD mouse model. In this study, we investigated the effect of atRA on LCWE-induced CA stenosis in a mouse model of KD.

Preliminary evaluation of LCWE-induced CA stenosis
The mice were sacri ced 2, 4, 8 and 16 weeks after LCWE injection to investigate the natural history of LCWE-induced CA stenosis. Elastica van Gieson (EVG) staining revealed that CA intimal formation was rst observed at 2 weeks and that the intimal thickness gradually increased over time. In addition to the increased intimal thickness, maximum vessel luminal narrowing was observed at 16 weeks after LCWE administration (Fig 1). Therefore, we chose to begin atRA treatment 2 weeks after LCWE injection and continued treatment for the next 14 weeks, which coincided with the onset of intimal formation and the time of maximum CA stenosis in this mouse model. None of the control mice that were injected with phosphate-buffered saline (PBS) exhibited pathological changes (data not shown).
Effect of atRA on CA in ammation and stenosis Next, we evaluated the effects of atRA on CA in ammation. atRA (20 mg/kg) was orally administered 5 days per week from 2 to 16 weeks after LCWE injection. In ammatory cells predominantly in ltrated the aortic root, and bilateral CAs were observed in LCWE-induced mice compared to PBS-treated mice. atRA signi cantly suppressed CA in ammation (19.3±2.8 vs 4.4±2.8, p<0.0001) (Fig. 2). Mice stimulated with LCWE exhibited CA stenosis in addition to vasculitis. We next assessed the effects of atRA on CA stenosis using three parameters. Representative microphotographs showing LCWE-induced CA intimal formation are shown in Fig. 3a. atRA signi cantly reduced intimal incidence (100% vs 18.5%, p<0.05), intimal thickness (100.5±18 vs 11.5±9.3 μm, p<0.01), and the CA stenosis rate (67.5 vs 7.6%, p<0.01) (

Effect of atRA on LCWE-induced elastin degradation through the suppression of MMP-9
Changes in the proliferative phenotype of SMCs precede elastolysis and are thought to play an important role in the development of intimal hyperplasia 20 . Therefore, assessing elastin degradation is extremely important for regulating intimal formation in the vessel wall. Next, we investigated the effect of atRA on the frequency of elastic breaks in the tunica media. LCWE-stimulated mice had more frequent interruptions and weakening of elastic bers than mice that were administered PBS (Fig. 5a). atRA signi cantly reduced the elastin break scores of the external elastic lumina (EEL) (28±1 vs 6.9±3.4, p<0.0001) and internal elastic lumina (IEL) (21.2±1.7 vs 3.6±2.1, p<0.0001) (Fig. 5b,c). The potent electrolytic protein MMP-9, was increased in the serum of LCWE-induced mice (1.226±0.18 ng/ml) compared with PBS-injected mice (0.697±0.12 ng/ml, p=0.66). This LCWE-induced increase in MMP-9 was signi cantly suppressed in atRA-treated mice (0.674±0.12 ng/ml, p=0.035 vs LCWE group) (Fig. 5d).

Inhibitory effects of atRA on SMC migration in vitro
Next, human coronary artery smooth muscle cells (HCASMCs) were used to investigate the effect of atRA on HCASMC migration (Fig. 6a). The cell migration assay revealed that platelet-derived growth factor subunit B homodimer (PDGF-BB) stimulation increased the area covered by migrated cells (n=16, 515,703 μm 2 ) compared to that of medium alone (n=8, 443,594 μm 2 , p=0.04). Cells were then treated with 0.1, 1.0, and 10 nM atRA for 72 h. The areas covered by migrated cells after treatment with 0.1 and 1 nM atRA were 407,610 (n=16) and 424,162 μm 2 (n=16), respectively. While these concentrations of atRA induced signi cant reductions of 21% (p<0.0001) and 18% (p=0.002), respectively, compared to those in the PDGF-BB-treated group, a greater reduction of 49% was observed in the 10 nM atRA treatment group (n=16, p<0.0001 vs. PDGF-BB, 0.1, 1 nM atRA treatment) (Fig. 6b).

Discussion
In the present study, we found that atRA dramatically reduced intimal hyperplasia and alleviated CA stenosis in an LCWE-induced model of KD vasculitis. CA stenosis is clinically caused by thrombus formation or intimal hyperplasia and induces cardiac events such as cardiac ischemia, MI, and even sudden death 2,4 . Vascular smooth muscle proliferation plays a pivotal role in the development of intimal hypertrophy, causing CA stenosis in KD patients with CAA, as evidenced by autopsy studies 9 . Therefore, this is the rst report focused on the prevention of CA stenosis by regulating the properties of SMCs in a mouse CA arteritis model. atRA is the most active metabolite of vitamin A. Numerous studies have reported that atRA has biological effects on various types of tumors, including breast and lung cancer and APL 21 . In recent years, atRA has been used as the standard therapeutic drug for the treatment of adult APL and pediatric neuroblastoma 22 . On the other hand, several basic experimental studies of cardiovascular disorders have shown that atRA has antiproliferative and antimigratory effects in animal models of intimal hyperplasia. Miano et al. showed that atRA reduced neointimal formation and promoted favorable geometric remodeling of the rat carotid artery after balloon withdrawal injury 18 . In addition, Zhang et al. showed that atRA suppressed neointimal hyperplasia and inhibited VSMC proliferation and migration through direct activation of AMP-activated protein kinase (AMPK) and inhibition of mTOR signaling 23 . Therefore, we hypothesized that atRA might exert bene cial effects on the CA stenosis mouse model we developed in recent years. These previous data indicated that atRA improved intimal proliferation mainly associated with αSMA-positive cells. However, the effectiveness of atRA on cardiovascular disorders in clinical practice has not yet been veri ed. In addition, several clinical studies have investigated the relationship between retinol binding protein 4 (RBP4) and KD. Kimura et al. showed that RBP4, which is a candidate diagnostic marker, was decreased in patients with acute KD 24 . Recently, Yang et al. reported that KD patients had signi cantly lower RBP4 levels than healthy controls, suggesting that RBP4, which is a main retinol transport protein, is closely associated with markers of in ammation and thrombogenesis in children with KD 25,26 .
Notably, compared to untreated mice, mice treated with atRA had signi cantly reduced CA in ammatory scores. This anti-in ammatory effect was consistent with the data reported by Miyabe et al. Am80, a retinoic acid receptor (RAR) agonist, has been shown to ameliorate mouse vasculitis induced by CAWS by suppressing neutrophil migration and activation 19 . In our study, the underlying pathophysiological mechanism of the anti-in ammatory effect of atRA remains unclear, but it is hypothesized that the SMC phenotype predisposes patients to increased proliferation and migration and contributes to persistent in ammation of the vessel wall.
We found that atRA signi cantly decreased elastin breaks and suppressed serum MMP-9 activity. A previous study by Axel et 20 . Therefore, it is reasonable to hypothesized that atRA protects against elastin degradation through the downregulation of MMP-9 activity, which in turn results in the suppression of proliferative phenotypic switching and the inhibition of intimal hyperplasia.
In vitro, we showed direct inhibitory effects of atRA on migration using HCASMCs stimulated with PDGF-BB. Several in vivo and in vitro studies have investigated the pathways that regulate the migration or proliferation of VSMCs. Day et al. rst reported that atRA inhibited airway SMC migration by modulating the phosphatidylinositol 3 kinase (PI3K)/Akt pathway 29 . In addition, Zhang et al. demonstrated that atRA might inhibit neointimal hyperplasia and suppress VSMC proliferation and migration by direct activation of AMP-activated protein kinase (AMPK). They concluded that AMPK might be the pharmacological target of ATRA and that activation of AMPK by atRA may be a novel treatment strategy for atherosclerosis 23 . More recently, Yu et al. reported that atRA prevented vein graft stenosis by inhibiting Rb-E2F-mediated cell cycle progression in human vein SMCs 30 . It was also reported that the Rb-E2F pathway was required for PDGF-BB-induced VSMC proliferation 31 . Our results con rmed that atRA inhibited PDGF-BB-induced HCASMC migration, suggesting an association between the Rb-E2F pathway and the antiproliferative effect of atRA.
Our research has some limitations. First, it was not possible to clearly determine whether atRA had a signi cant effect on in ammatory suppression or the induction of vascular repair. This is because we did not observe pathological features or changes in proin ammatory cytokines and chemokines over time. One possibility remains that atRA provides complete protection from the development of CA in ammation and stenosis during the experimental period. Treatment with atRA was started relatively early in this mouse model. Therefore, it was considered necessary to select a schedule for later administration of atRA treatment when CA in ammation had already developed. Second, the therapeutic goal of KD patients with CAA is to not only promote the regression of CA aneurysms but also prevent further cardiac events such as acute MI and sudden death. Thus, the bene cial effect of atRA on the suppression of intimal hyperplasia may lead to protection against CA stenosis and these harmful events. On the other hand, excessive inhibition of intimal hyperplasia may delay aneurysmal regression, resulting in a residual aneurysm. Therefore, it is important to induce favorable SMC proliferation rather than completely suppressing intimal hyperplasia leading to CA stenosis. Additional research is needed to elucidate the mechanism of the bene cial effects of atRA.
In conclusion, atRA dramatically reduced CA in ammation and stenosis by suppressing the production of MMP-9 and the migratory properties of SMCs by regulating cellular functions. Therefore, atRA, which has both anti-in ammatory effects and the ability to repair the vascular wall, is expected to have clinical applications to prevent CA stenosis in KD patients with CAA.

Experimental protocol
Five-week-old male C57BL/6J mice were purchased from CLEA Japan (Tokyo, Japan) and maintained in an environment with a 12 h light/12 h dark cycle under speci c pathogen-free conditions. First, we examined the natural course of coronary stenosis in mice after LCWE administration. The mice were sacri ced on weeks 2, 4, 8, and 16 after intraperitoneal injection of 1,000 µg of LCWE (n=5 to 7 in each group). Cardiac tissue was harvested, and histological assessments were performed. Based on the pathological features of CA stenosis observed in the preliminary natural course experiment, the effects of atRA on CA stenosis in this mouse model of KD were examined. Five-week-old male C57BL/6J mice (n=13) were intraperitoneally injected with 1,000 μg of LCWE. Two weeks later, these mice were divided into two groups. Mice were orally administered atRA (Sigma-Aldrich, St. Louis, MO, USA) at a dose of 20 mg/kg (n=7) or an equivalent volume (0.2 ml) of corn oil (n=6) using a disposable exible-type gastric tube (Fuchigami, Kyoto, Japan) 5 days per week for 16 weeks. The same dose of atRA has been shown previously to inhibit neointimal hyperplasia and suppress the proliferation and migration of VSMCs in mice 23 . Control mice (n=5) were injected with PBS instead of LCWE. All mice were killed at 18 weeks after LCWE or PBS administration. All mouse experimental procedures were performed in accordance with institutional guidelines and regulations of Saitama Children's Medical Center, Japan. This animal experiment was performed with the approval of the Animal Experimental Ethics Committee of Saitama Children's Medical Center (No: 2020-003). All experiments were performed in accordance with relevant ARRIVE guidelines.
LCWE preparation LCWE was prepared as previously described 12,32 . Brie y, Lactobacillus Paracasei subsp. paracasei (ATCC 11578; American Type Culture Collection, Manassas, VA, USA) was cultured in MRS broth (BD Difco, Franklin Lakes, NJ, USA) for 48 h at 37 °C. The cells were harvested and washed with PBS, after which the cells were disrupted in 2 packed volumes of 4% sodium dodecyl sulfate (SDS) overnight at room temperature. Cell wall fragments were extensively washed 10 times with PBS to remove any residual SDS. The SDS-treated cell wall fragments were sonicated (5 g of packed wet weight in 15 ml of PBS) for 2 h using a Q500 sonicator with a 3/4" diameter probe at an amplitude setting of 70% (QSonica LLC, CT, USA). During sonication, the cell wall fragments were maintained by cooling in a dry ice/ethanol bath. The supernatant was centrifuged for 1 h at 20,000 g at 4 °C, and the supernatant containing the cell wall extract was used for the experiments. The concentration of LCWE was determined based on the rhamnose content as measured by a phenol-sulfuric acid colorimetric assay and adjusted to 5,000 µg/ml. To induce coronary arteritis/stenosis in mice, 1,000 µg (0.2 ml) of the LCWE preparation was injected intraperitoneally.

Histological evaluation
An in ammatory assessment of the CAs was performed as described previously 12 . Brie y, 2.5-μm sections of cardiac tissue were stained with hematoxylin and eosin (H&E) and EVG. Then, we selected 5 consecutive sections slightly distal to the bifurcation of the bilateral CAs. The intensity of coronary arteritis was scored with the following four criteria: 0, no in ammation; 1, in ammatory cells in only the adventitia; 2, in ammatory cells in both the intima and adventitia; and 3, panvasculitis as previously described 33 . The intensity of coronary in ammation was the total score of the bilateral CAs and is expressed as the in ammatory score of 10 CAs per individual animal. Elastin breaks were scored on the following scale: score 0, no interruption of elastic bers; score 1, elastic breaks ≤10; 2, elastic breaks ≻10; and score 3, obscuration or disappearance of elastic bers. Elastin breaks were de ned as interruptions in the elastin ber and the reappearance of the ber and are expressed as the total number of breaks in ve consecutive sections. The IEL and EEL of the bilateral CAs were evaluated. In addition, coronary stenosis was assessed according to three different parameters, including the frequency of neointimal formation, intimal thickness (μm), and coronary stenosis rate, which were measured with an NIS-Elements AR Ver5.11.00 (Nikon Instruments, Inc, Tokyo, Japan).
Cell culture and migration assay Primary HCASMCs were purchased from Thermo Fisher Scienti c K.K. (Tokyo, Japan) These cells were cultured in Dulbecco's modi ed Eagle's medium (3.15 g/L glucose and 15 mM L-glutamine DMEM) (Lonza, Tokyo, Japan) containing 10% fetal bovine serum (FBS) and a penicillin-streptomycinamphotericin B mixture (Lonza, Tokyo, Japan) at 37 °C in a humidi ed atmosphere containing 5% CO 2 .
The culture medium was changed every 24 h. Once the HCASMCs had grown to approximately 80 to 90% con uence, the cells were detached with a trypsin EDTA solution (Thermo Fisher Scienti c, Tokyo, Japan). HCASMC migration assays were performed using an Oris cell migration assay kit (Platypus Technologies, LCC, WI, USA) according to the manufacturer's instructions. Brie y, cells were seeded at 10,000 cells/well in 96-well plates, and each well was coated with collagen I with a stopper in the central area to prevent the cells from adhering to the detection zone; the cells were incubated in serum-free medium for 24 h. Immediately after the stopper was removed, the cells were cultured with 20 ng/ml PDGF-BB in the absence or presence of various concentrations of atRA (0.1, 1.0, and 10 nM). After 72 h of incubation, cells that migrated from the perimeter to the detection zone were measured using an IN Cell Analyzer 2200 (Cytiba, MA, USA). The number of migrated cells in each well was counted in low-power (4×) elds.

Measurement of total MMP-9 activity
Blood samples were extracted directly from the left ventricle using a 1 ml syringe with a 27-gauge needle immediately before sacri ce under deep anesthesia. The blood samples were centrifuged at 1,500 rpm at room temperature and stored at -80 °C until use. The serum MMP-9 level was measured by an MMP-9 activity assay kit for mice (Cosmo Bio Co., Ltd, Tokyo, Japan) according to the manufacturer's instructions, and the data are expressed as ng/ml.

Statistical analysis
All values are presented as the mean ± standard error of the mean (SEM). Statistically signi cant differences between mean values were determined using a two-tailed Mann-Whitney test. Statistical differences among the three or more groups were determined by one-way ANOVA followed by Bonferroni post hoc test. Differences in which p<0.05 were considered statistically signi cant. IBM SPSS Statistics for Windows Version 24.0 (SPSS Japan, Tokyo, Japan) was used to analyze the data.