Legumain Promotes Atherosclerotic Vascular Remodeling

Legumain, a recently discovered cysteine protease, is increased in both carotid plaques and plasma of patients with carotid atherosclerosis. Legumain increases the migration of human monocytes and human umbilical vein endothelial cells (HUVECs). However, the causal relationship between legumain and atherosclerosis formation is not clear. We assessed the expression of legumain in aortic atheromatous plaques and after wire-injury-induced femoral artery neointimal thickening and investigated the effect of chronic legumain infusion on atherogenesis in Apoe−/− mice. We also investigated the associated cellular and molecular mechanisms in vitro, by assessing the effects of legumain on inflammatory responses in HUVECs and THP-1 monocyte-derived macrophages; macrophage foam cell formation; and migration, proliferation, and extracellular matrix protein expression in human aortic smooth muscle cells (HASMCs). Legumain was expressed at high levels in atheromatous plaques and wire injury-induced neointimal lesions in Apoe−/− mice. Legumain was also expressed abundantly in THP-1 monocytes, THP-1 monocyte-derived macrophages, HASMCs, and HUVECs. Legumain suppressed lipopolysaccharide-induced mRNA expression of vascular cell adhesion molecule-1 (VCAM1), but potentiated the expression of interleukin-6 (IL6) and E-selectin (SELE) in HUVECs. Legumain enhanced the inflammatory M1 phenotype and oxidized low-density lipoprotein-induced foam cell formation in macrophages. Legumain did not alter the proliferation or apoptosis of HASMCs, but it increased their migration. Moreover, legumain increased the expression of collagen-3, fibronectin, and elastin, but not collagen-1, in HASMCs. Chronic infusion of legumain into Apoe−/− mice potentiated the development of atherosclerotic lesions, accompanied by vascular remodeling, an increase in the number of macrophages and ASMCs, and increased collagen-3 expression in plaques. Our study provides the first evidence that legumain contributes to the induction of atherosclerotic vascular remodeling.


Introduction
Atherosclerosis is characterized by a complex process of vascular injury; inflammation, with monocyte adhesion to endothelial cells (ECs); lipid deposition within macrophage foam cells; neointimal hyperplasia, involving vascular smooth muscle cells (VSMCs); and extracellular matrix (ECM) remodeling [1,2]. Vascular inflammation is characterized by the up-regulation of tumor necrosis factor-α

Expression of Legumain in Mouse Atherosclerotic Lesions
Marked atheromatous plaques were observed in the aortas of 21-week-old Apoe −/− mice ( Figure 1A). Legumain was abundantly expressed in advanced and early-stage atheromatous plaques in the aortas of 21-week-old and 17-week-old Apoe −/− mice, respectively ( Figure 1B,C), but was not detected in the non-atherosclerotic aortas of 13-week-old control Apoe −/− mice ( Figure 1D).
Neointimal hyperplasia was observed in the obstructed left femoral arteries following wire injury in 15-week-old Apoe −/− mice ( Figure 1E). Abundant legumain expression was also detected in marked neointimal lesions in the injured left femoral arteries ( Figure 1F). In contrast, in the non-injured right femoral arteries ( Figure 1G), legumain was not expressed in the normal right femoral arteries of 15-week-old Apoe −/− mice ( Figure 1H). . Marked neointimal hyperplasia (reddish-purple areas surrounded by arrows) was observed in the obstructed left femoral artery following wire injury in 15-week-old Apoe −/− mice (E). In Apoe −/− mice, legumain (dark blown) was expressed at high levels in the advanced and early-stages of aortic atheromatous plaques at 21 weeks old (B) and 17 weeks old (C), respectively, and in left femoral artery neointimal lesions after wire injury at 15 weeks old (F). Legumain expression was not observed in the non-atherosclerotic aortas at 13 weeks old (D) or in the right normal femoral arteries at 15 weeks old (H). Each experiment was performed on at least two independent occasions to ensure reproducibility. AV = aortic valve, L = lumen.

Effects of Legumain on Atherosclerotic Lesion Development in Apoe −/− Mice
There were no significant differences in body weight; food intake; systolic or diastolic blood pressure; or fasting plasma levels of glucose, total cholesterol, triglycerides, free fatty acids, or insulin between Apoe −/− mice treated with saline and those treated with legumain (Table 1). However, chronic legumain infusion tended to aggravate atherosclerotic lesions on the aortic luminal surface ( Figure  2A,B,O), with somewhat increased plaque burden on the aortic wall and vascular inflammation (pentraxin-3 expression) ( Figure 2C,D,I,J,P,S); there were no significant differences. Legumain significantly increased the number of macrophages, VSMCs, and collagen fibers as well as collagen-3 expression in the aortic sinus wall ( Figure 2E-H,K-N,Q,R,T,U). Legumain tended to reduce the macrophage/VSMC ratio, which is a surrogate marker of atheromatous plaque instability ( Figure 2V).  Marked neointimal hyperplasia (reddish-purple areas surrounded by arrows) was observed in the obstructed left femoral artery following wire injury in 15-week-old Apoe −/− mice (E). In Apoe −/− mice, legumain (dark blown) was expressed at high levels in the advanced and early-stages of aortic atheromatous plaques at 21 weeks old (B) and 17 weeks old (C), respectively, and in left femoral artery neointimal lesions after wire injury at 15 weeks old (F). Legumain expression was not observed in the non-atherosclerotic aortas at 13 weeks old (D) or in the right normal femoral arteries at 15 weeks old (H). Each experiment was performed on at least two independent occasions to ensure reproducibility. AV = aortic valve, L = lumen.

Effects of Legumain on Atherosclerotic Lesion Development in Apoe −/− Mice
There were no significant differences in body weight; food intake; systolic or diastolic blood pressure; or fasting plasma levels of glucose, total cholesterol, triglycerides, free fatty acids, or insulin between Apoe −/− mice treated with saline and those treated with legumain (Table 1). However, chronic legumain infusion tended to aggravate atherosclerotic lesions on the aortic luminal surface (Figure 2A,B,O), with somewhat increased plaque burden on the aortic wall and vascular inflammation (pentraxin-3 expression) ( Figure 2C,D,I,J,P,S); there were no significant differences. Legumain significantly increased the number of macrophages, VSMCs, and collagen fibers as well as collagen-3 expression in the aortic sinus wall ( Figure 2E-H,K-N,Q,R,T,U). Legumain tended to reduce the macrophage/VSMC ratio, which is a surrogate marker of atheromatous plaque instability ( Figure 2V).

Effects of Legumain on Inflammatory Responses in HUVECs
Legumain had no significant effect on the mRNA expression of IL6, TNFA, ICAM1, VCAM1, or SELE, but LPS significantly stimulated the expression of these mRNAs in HUVECs ( Figure 4A,B,D-F). However, legumain tended to increase MCP1 mRNA expression to the same extent as LPS ( Figure 4C). In HUVECs, legumain significantly suppressed the LPS-induced mRNA expression of VCAM1 ( Figure 4E) and tended to decrease LPS-induced TNFA and MCP1 expression ( Figure 4B,C), but it did not affect LPS-induced ICAM1 expression ( Figure 4D). In contrast, legumain tended to increase the LPS-induced mRNA expression of IL6 and SELE in HUVECs ( Figure 4A,F); there were no significant differences.

Effects of Legumain on Inflammatory Responses in HUVECs
Legumain had no significant effect on the mRNA expression of IL6, TNFA, ICAM1, VCAM1, or SELE, but LPS significantly stimulated the expression of these mRNAs in HUVECs ( Figure 4A,B,D-F). However, legumain tended to increase MCP1 mRNA expression to the same extent as LPS ( Figure  4C). In HUVECs, legumain significantly suppressed the LPS-induced mRNA expression of VCAM1 ( Figure 4E) and tended to decrease LPS-induced TNFA and MCP1 expression ( Figure 4B,C), but it did not affect LPS-induced ICAM1 expression ( Figure 4D). In contrast, legumain tended to increase the LPS-induced mRNA expression of IL6 and SELE in HUVECs ( Figure 4A,F); there were no significant differences.

Effects of Legumain on Inflammatory Responses in HUVECs
Legumain had no significant effect on the mRNA expression of IL6, TNFA, ICAM1, VCAM1, or SELE, but LPS significantly stimulated the expression of these mRNAs in HUVECs ( Figure 4A,B,D-F). However, legumain tended to increase MCP1 mRNA expression to the same extent as LPS ( Figure  4C). In HUVECs, legumain significantly suppressed the LPS-induced mRNA expression of VCAM1 ( Figure 4E) and tended to decrease LPS-induced TNFA and MCP1 expression ( Figure 4B,C), but it did not affect LPS-induced ICAM1 expression ( Figure 4D). In contrast, legumain tended to increase the LPS-induced mRNA expression of IL6 and SELE in HUVECs ( Figure 4A,F); there were no significant differences.

Effects of Legumain on the Inflammatory Phenotype of Human Macrophages
After 6 days of culture, the differentiation of THP-1 monocytes to macrophages was confirmed by increased expression of CD68, a macrophage differentiation marker ( Figure 5A,B). Legumain did not affect monocyte differentiation to macrophages. However, legumain (50 ng/mL) increased the expression of macrophage receptor with collagenous structure (MARCO), an M1 macrophage marker, but not arginase-1, an M2 macrophage marker, during differentiation ( Figure 5A,B).

Effects of Legumain on the Inflammatory Phenotype of Human Macrophages
After 6 days of culture, the differentiation of THP-1 monocytes to macrophages was confirmed by increased expression of CD68, a macrophage differentiation marker ( Figure 5A,B). Legumain did not affect monocyte differentiation to macrophages. However, legumain (50 ng/mL) increased the expression of macrophage receptor with collagenous structure (MARCO), an M1 macrophage marker, but not arginase-1, an M2 macrophage marker, during differentiation ( Figure 5A,B).

Effects of Legumain on Foam Cell Formation
Treatment of THP-1 monocyte-derived macrophages with OxLDL significantly increased foam cell formation ( Figure 6A). Further, legumain (10,50 ng/mL) significantly enhanced oxLDL-induced foam cell formation ( Figure 6A). Legumain significantly increased the protein expression of SR-A, ACAT-1, and NCEH in a concentration-dependent manner ( Figure 6C-E) but did not significantly alter the expression of CD36 or ABCA1 ( Figure 6B,F).

Effects of Legumain on Foam Cell Formation
Treatment of THP-1 monocyte-derived macrophages with OxLDL significantly increased foam cell formation ( Figure 6A). Further, legumain (10, 50 ng/mL) significantly enhanced oxLDL-induced foam cell formation ( Figure 6A). Legumain significantly increased the protein expression of SR-A, ACAT-1, and NCEH in a concentration-dependent manner ( Figure 6C-E) but did not significantly alter the expression of CD36 or ABCA1 ( Figure 6B,F).

Effects of Legumain on Migration, Proliferation, and Apoptosis of HASMCs
Legumain (25 ng/mL) significantly increased the migration of HASMCs (p < 0.0001, Figure 7A). However, legumain did not significantly affect the proliferation or apoptosis of HASMCs ( Figure  7B,C).

Effects of Legumain on Migration, Proliferation, and Apoptosis of HASMCs
Legumain (25 ng/mL) significantly increased the migration of HASMCs (p < 0.0001, Figure 7A). However, legumain did not significantly affect the proliferation or apoptosis of HASMCs ( Figure 7B,C).

Discussion
This is the first demonstration that legumain, a newly discovered cysteine protease, potentiates atherosclerotic vascular remodeling. Legumain enhances macrophage foam cell formation and VSMC migration; increases collagen-3, fibronectin, and elastin expression in VSMCs in vitro; and induces the formation of atherosclerotic lesions, with increased numbers of macrophages and VSMCs and increased collagen-3 expression in Apoe −/− mice in vivo. The effects of legumain on atherogenesis are not as strong as the effects of other vasoactive agents [30,31]. However, it is worth noting that legumain did promote atherosclerotic vascular remodeling, associated with increased ECM production via the PI3K/Akt pathway, in VSMCs.
Previous studies have shown that legumain regulates ECM remodeling by inducing the degradation of collagen-1 and fibronectin [23,32] and exerts anti-fibrotic effects by attenuating the deposition of collagen and fibronectin in a mouse model of obstructive nephropathy [33]. A recent study has shown that legumain increases the synthesis of ECM proteins by activating MMP-2/transforming growth factor-β1 signaling in pulmonary artery VSMCs in a mouse model of pulmonary hypertension [24]. When considered in integration, legumain may contribute to tissue remodeling by repeating alternately the scrap and build of ECM under various pathophysiological conditions. In addition, a previous study has shown that the expression levels of legumain are higher in unstable plaques than in stable plaques in human carotid arteries [27]. Because legumain is co-expressed with MMP, it is regarded as a contributor to plaque rupture [27]. In our study, legumain increased the number of intraplaque collagen fibers and increased the expression of collagen-3, fibronectin, and elastin in VSMCs. Moreover, it decreased the macrophage/VSMC ratio, which is a marker of plaque instability. Our findings suggest that legumain may play a key role in the elasticity and stabilization of atheromatous plaques. Further studies are needed to determine whether legumain can stabilize plaques at the advanced atherosclerotic phase in Apoe −/− mice.
The trafficking of legumain in and outside the cell has been reported by Dall E and Brandstetter H [14,34]. Prolegumain is translocated via the endoplasmic reticulum and Golgi apparatus to the endo-lysosomal system, where it is activated to legumain as asparaginyl endopeptidase with acidic pH [34]. Endo-lysosomal pH is decreased by IL-6 and increased by IL-10, leading to the activation and inactivation of legumain, respectively [14]. Legumain may be extracellularly secreted directly via the Golgi apparatus or indirectly via the endosomal system. Extracellular legumain may re-enter the cell via endocytosis or directly via translocation to the cytoplasm [14]. However, the receptors for legumain have not been identified so far. These findings suggest that legumain may act in an autocrine/paracrine manner. Therefore, legumain is also regarded as a kind of local and systemic hormone.
Several studies have shown that plasma concentrations of legumain are~0.6 and~1.5 ng/mL in healthy subjects [22,24] and~2.0 ng/mL in patients with carotid atherosclerosis [22]. The concentrations of legumain required to elicit THP-1 monocyte-derived macrophage, HUVEC, and HASMC responses in the present study were 7.5-75 ng/mL. Local levels of vasoactive agents, produced by vascular cells, may increase to similar concentrations and act in an autocrine/paracrine manner [35,36]. In addition, different concentrations of legumain were required to induce foam cell formation and related protein expression in THP-1 monocyte-derived macrophages in the present study. This was mostly dependent on the presence or absence of oxLDL. There were also differences in the adequate concentrations of legumain required to elicit different responses among different cells. Cell type-specific signal transduction mechanisms may be responsible for these differences. Further, when the concentration of legumain is higher over the adequate one, it did not exert cytotoxicity. Cells treated with high concentrations of legumain were intact, which were demonstrated by morphological observation, cell variability, and housekeeping gene expression.
There are some limitations in considering the results from animal experiments. The number of control group was low, because some mice infused with saline died by cannibalism. It was possible that small sample size comparisons did not lead to the statistically significant differences in atherosclerosis and vascular inflammation and remodeling. In addition, the knockout of legumain and administration of legumain inhibitor or neutralizing antibody in Apoe −/− mice may strengthen to demonstrate the stimulatory effects of legumain on atherosclerotic vascular remodeling. Likewise, the examination using legumain with its inhibitors or an inactive form of legumain, prolegumain, may be required in all in vitro experiments in future studies.
The present study gives new insights into the role of legumain in the pathophysiology of atherosclerosis. These results have important clinical implications, identifying legumain as a potential novel target for the treatment of atherosclerosis. Low molecular weight inhibitors and neutralizing antibodies against legumain may become promising candidate drugs for the treatment of atherosclerosis. Several studies have shown that legumain inhibitors suppress cancer metastasis and the pathogenesis of Alzheimer's disease [14,37,38]. In addition, recent studies have reported that statins (atorvastatin and simvastatin) suppress the expression and activity of legumain in human monocytes/macrophages and in human myotubes [21,39,40]. These interventions against legumain may provide effective therapeutic strategies for preventing atherosclerosis and related vascular remodeling.
In conclusion, the results from the present study indicated that legumain promoted atherosclerotic vascular remodeling by enhancing macrophage foam cell formation; VSMC migration; and collagen-3, fibronectin, and elastin production by VSMCs. Legumain is expected to emerge as a novel therapeutic target for atherosclerosis and related diseases.

Materials
Human recombinant legumain was purchased from Raybiotech (Norcross, GA, USA) for in vitro experiments and from Cusabio (Houston, TX, USA) for in vivo experiments, with the purity of 95% and 90%, respectively. A rabbit polyclonal antibody against human legumain was purchased from Bioss (Woburn, MA, USA). LPS and phorbol 12-myristate 13-acetate were purchased from Sigma (St. Louis, MO, USA) and Wako (Osaka, Japan), respectively. HUVECs and HASMCs were purchased from Lonza (Basel, Switzerland) and THP-1 monocytes were from the Health Science Research Resources Bank (Osaka, Japan).

Administration of Legumain to Mice
Animal experiments were performed in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals, with protocols approved by the Institutional Animal Care and Use Committee of Tokyo University of Pharmacy and Life Sciences (No. L18-02). In this study, spontaneously hyperlipidemic Apoe −/− mice were used as an animal model for atherosclerosis. The Apoe −/− mice display poor lipoprotein clearance with subsequent accumulation of cholesterol ester-enriched particles in the blood, which promote the development of atherosclerotic plaques.
A total of 16 male Apoe −/− mice (BALB/c. KOR/StmSlc-Apoe shl mice), at the age of 9 weeks, were purchased from Japan SLC (Hamamatsu, Japan) and maintained on a normal diet until 13 weeks of age, followed by a high-cholesterol diet (Oriental Yeast, Tokyo, Japan) until the experimental endpoint. At 13 and 17 weeks of age, 4 mice with no or minor atherosclerotic lesions in the aorta, respectively, were sacrificed as controls. The remaining 12 mice at 17 weeks of age were divided into 2 groups and these were administered saline (vehicle, n = 4) or legumain (5 µg/kg/h, n = 8) for 4 weeks using osmotic mini-pumps (Alzet Model 1002; Durect, Cupertino, CA, USA). Once every 2 weeks, the mini-pumps were implanted subcutaneously into the dorsum under medetomidine/midazolam/butorphanol anesthesia [2,5].

Assessment of Legumain Expression in Intimal Lesions following Wire Injury
Under deep anesthesia, vascular injury was induced with a spring guide wire (diameter: 0.014 inches) in the left femoral artery of 12-week-old Apoe −/− mice fed a high-cholesterol diet [10]. Three weeks after the procedure, the mice were sacrificed to determine arterial neointimal thickness [10]. Paraffin-embedded cross-sections of wire-injured and non-injured femoral arteries were stained with Elastica-Van Gieson and an anti-legumain antibody.

Statistical Analysis
All values are expressed as means ± SEM. Data were analyzed by unpaired Student's t-test for 2 groups and by one-way analysis of variance, followed by Bonferroni's post hoc test, for ≥3 groups, using Statview-J 5.0 (SAS Institute, Cary, NC, USA). Statistical significance was defined as p < 0.05.