EPA Prevents the Development of Abdominal Aortic Aneurysms through Gpr-120/Ffar-4

Abdominal aortic aneurysms (AAAs), which commonly occur among elderly individuals, are accompanied by a risk of rupture with a high mortality rate. Although eicosapentaenoic acid (EPA) has been reported to prevent AAA formation, the mechanism by which EPA works on vascular smooth muscle cells is unknown. This study aimed to investigate the mechanism by which orally-administered EPA prevents the formation of severe AAAs that develop in Osteoprotegerin (Opg) knockout (KO) mice. In the CaCl2-induced AAA model, EPA attenuated the enhanced progression of AAAs in Opg-KO mice, including the increase in aortic diameter with destruction of elastic fibers in the media. Immunohistochemical analyses showed that EPA reduced the phosphorylation of transforming growth factor beta-activated kinase-1/Map3k7 (Tak-1) and c-Jun NH2-terminal kinase (JNK), as well as the expression of Matrix metalloproteinase-9 (Mmp-9) in the media of the aorta. In smooth muscle cell cultures, rh-TRAIL-induced activation of the Tak-1-JNK pathway and increase in Mmp-9 expression were inhibited by EPA. Moreover, GW9508, a specific ligand for G-protein coupled receptor (Gpr)-120/Free fatty acid receptor (Ffar)-4, mimicked the effects of EPA. The effects of EPA were abrogated by knockdown of the Gpr-120/Ffar-4 receptor gene. Our data demonstrate that the Trail-Tak-1-JNK-Mmp-9 pathway is responsible for the enhancement of AAAs in Opg-KO mice, and that EPA inhibits the Tak-1-JNK pathway by activating Gpr-120/Ffar-4, which results in the attenuation of AAA development.


Introduction
An abdominal aortic aneurysm (AAA) is a condition in which the abdominal aorta is expanded, leading to rupture with a high mortality rate. Destruction of medial elastic fibers and infiltration of macrophages into AAA lesions are considered as the main causes of AAA formation [1]. Degradation of aortic medial elastic fibers may be induced by imbalances between the proteolytic activities of matrix metalloproteinases, such as MMP-2 and MMP-9, and their inhibitors, the tissue inhibitors of metalloproteinases (Timp-1). The expression of both is reported to be elevated during aneurysm development [2][3][4][5][6]. It has been reported that MMPs are induced by activation of the JNK pathway [5,7,8] in macrophages and vascular smooth muscle cells (SMCs), and that inhibition of the JNK pathway results in AAA regression [7]. Pro-inflammatory cytokines, such as tumor necrosis factor (TNF)-α and interleukin (IL)-1, are known to activate the JNK pathway [9].
OPG, which is secreted from osteoblasts, promotes bone formation by acting as a decoy receptor for the receptor activator of nuclear factor kappa-B ligand (Rankl), a factor that promotes the differentiation and activation of osteoclasts during osteogenesis [10,11]. OPG is expressed in vascular SMCs, and plays roles in preventing arterial calcification and stabilizing plaque formation in animal models [12][13][14]. We recently reported that Opg prevents the development of AAAs, given the reduced formation of AAAs with complete disruption of medial elastic fibers in Opg-KO mice [15]. TNF-related apoptosis inducing ligand (Trail), which was increased in Opg-KO mice, induced the expression of Mmp-9 through JNK, and of Timp-1 through nuclear factor (NF)-κB pathways, in aortic SMCs. Since OPG functions as a decoy receptor for Trail, we concluded that Opg prevents AAA formation through its antagonistic effect on Trail.
Eicosapentaenoic acid (EPA), an omega-3 polyunsaturated fatty acid (ω-3 PUFA), showed great promise in the prevention of cardiovascular diseases through its anti-inflammatory effects [16][17][18]. In particular, EPA attenuated the development of AAAs through suppression of tissue remodeling and inhibition of macrophage-mediated inflammation [19] [20]. The molecular mechanism underlying EPA's anti-inflammatory effects in the vascular system is not yet fully understood. Recently, EPA has been reported to function as a ligand for Gpr-120/ Ffar-4. Activation of Gpr-120/Ffar-4 was shown to induce the release of glucagon-like peptide (GLP)-1 into the intestine [21], and also exhibited a possible anti-inflammatory effect by inhibiting JNK and NF-κB pathways in macrophages and adipocytes [22]. It is unknown whether Gpr-120/Ffar-4 is expressed in vascular SMCs, and whether it plays a role in AAA formation.
Here, we report that orally-administered EPA prevents the enlargement of AAAs in Opg-KO mice in a CaCl 2 -induced AAA model. We found the size of the abdominal aorta to be significantly reduced in Opg-KO mice fed an EPA (+) diet. In vascular SMCs, EPA and GW9508, agonists of Gpr-120/Ffar-4, significantly reduced the TRAIL-induced expression of Mmp-9 by inhibiting the JNK pathway. This suppression was reversed by knock-down of Gpr-120/Ffar-4 expression. Our findings suggest that EPA can prevent the enlargement of AAAs by activating Gpr-120/Ffar-4-mediated signaling in aortic SMCs.

Generation of a Mouse Model for AAAs
The experiment protocol was approved by the Committee of Animal Experimentation at Hiroshima University (A08-32) and carried out in accordance with this protocol. All surgeries were performed under sodium phenobarbital anesthesia, and efforts were made to minimize suffering during and after surgery. Animals were sacrificed by anesthetic overdose with sodium phenobarbital. Wild-type and Opg-KO male mice of the C57BL/6J strain (CLEA Japan, Inc.) were used. AAA was induced by the CaCl 2 method as described in our previous report [15]. Briefly, we added 50 μL of 0.5 M CaCl 2 to a small piece of cotton and placed it on the aortic wall for 15 min. The CaCl 2 -treated area was limited to 1/2 to 1/3 of the circumference surface. Characteristic of our peri-aortic CaCl 2 application is minimized invasion of the aorta. The F1-diet, which does not contain fish meat (Funabashi Farm Co., Ltd.), was mixed 5% (wt/wt) with EPA (provided by Mochida Pharmaceutical Co., Ltd.) and fed to mice starting two weeks before AAA induction up to the point of aorta examination.

Morphological Examination
Paraffin-embedded aortic tissues were used to generate 6 μm-thick sections that were subsequently subjected to hematoxylin & eosin (H&E), Elastica van Gieson (EVG), Aniline Blue-Azan, or von Kossa staining using standard protocols. The maximum aortic diameter, including thickened adventitia with inflammation, was measured as the external aortic diameter. For widths of the medial layers of the aorta, the maximum values were measured on the sides to which CaCl 2 was applied. The means of maximum and minimum measurements of the internal cross-sections of the aorta reflect the internal diameter of the aorta. All measurements were performed using Photoshop (Adobe Systems) and Image-J (National Institutes of Health; NIH).

Cell Culture
Mouse aortic SMCs were isolated from the abdominal aorta (from the diaphragm to the bifurcation) in five-week-old wild-type male mice, as previously described [15,23], and maintained in Dulbecco's Modified Eagle Medium (DMEM) with 10-20% fetal bovine serum (FBS). Cells were growth arrested at 70-80% confluence via incubation in serum-free medium for 48 hours before stimulating cells with rh-TRAIL (CD253; AbD Serotec). Different concentrations of EPA sodium salt (Sigma Aldrich) in 1% fatty acid-free bovine serum albumin (BSA; Calbiochem) in serum-free medium were applied to cells two hours before rh-TRAIL induction. Different concentrations of GW9508 (Cayman chemical) for activation of Gpr-120/Ffar-4, and either AS-601245 (Enzo Life Sciences) or SP600125 (Calbiochem), for inhibition of JNK activity, were added two hours before rh-TRAIL induction. Gpr-120/Ffar-4 siRNA (Ambion silencer select) for knock-down was applied 24 hours before rh-TRAIL induction.

Western Blotting
Cell lysates were prepared from growth-arrested vascular SMCs (VSMCs), which were stimulated with rh-TRAIL and then collected in radioimmunoprecipitation assay (RIPA) lysis buffer by scraping. Samples containing equal amounts of proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a nitrocellulose membrane. The blot was incubated with a primary antibody (Phospho-SAPK/JNK, rabbit monoclonal, 1:1000, Cell Signaling; SAPK/JNK, rabbit monoclonal, 1:1000, Cell Signaling; Phospho-TAK-1, rabbit monoclonal, 1:1000, Cell Signaling; TAK-1, rabbit monoclonal, 1:1000, Cell Signaling) and subsequently with an appropriate secondary antibody. Signals were visualized with the enhanced chemiluminescence (ECL) system (Amersham-Pharmacia Co, U. K.). Images were captured using Photoshop (Adobe) and densitometry was performed using Image J software (NIH).

Statistical Analysis
All data were statistically evaluated using Ekuseru-Toukei 2012 software (Social Survey Research Information Co., Ltd.) and quantitative values are expressed as mean ± standard deviation (SD). For mouse studies, a non-parametric analysis, Kruskal-Wallis test with Scheffe's post-hoc analysis or Steel-Dwass post-hoc analysis, was performed. For cell culture experiments, which were performed independently three times, the Mann-Whitney U test was used for comparisons. For all statistical tests, P<0.05 was considered significant.

EPA Attenuates the Development of AAAs in Opg-KO Mice
In order to examine whether EPA attenuates the development of AAA in Opg-KO mice, we first measured the size of the aorta in wild-type and Opg-KO mice, which were fed a diet with (+) or without (-) EPA starting two weeks before AAA induction [7], [15]. In wild-type mice, regardless of feeding with EPA or not, a significant increase in external aortic diameters was observed at six weeks after AAA induction (Fig 1A and 1B). No significant difference was found in external aortic diameters with or without EPA. In Opg-KO mice, the external aortic diameter was increased at one week. At six weeks, however, obvious enlargement of the external aortic diameter was observed in mice fed the EPA (-) diet, while this enlargement was significantly reduced in mice fed the EPA (+) diet (Fig 1A and 1B). No significant change in mouse body weight by EPA was observed (S1A Hematoxylin and eosin staining of aortic sections showed an increase of the internal diameter of the abdominal aorta and thinning of the medial layer at one week in wild-type and Opg-KO mice fed EPA (-) and (+) diets (Fig 1C-1E). Six weeks later, an increase of the internal diameter was maintained in both wild-type and Opg-KO mice. The width of the medial layer was markedly enlarged in Opg-KO mice fed the EPA (-) diet. This enlargement was significantly reduced with the EPA (+) diet (Fig 1C and 1E). Elastica van Gieson staining showed typical wavy lamellae of elastic fibers in the medial layers of wild-type and Opg-KO mice at day 0 ( Fig 1F). In Opg-KO mice fed the EPA (-) diet, disruption of the medial elastic lamellae was observed at one week after AAA induction; at six weeks, medial elastic lamellae were completely disrupted. However, in Opg-KO mice fed the EPA (+) diet, elastic fibers were retained with flattened structures (Fig 1F). This suggests that EPA attenuates the adverse effects of the loss-of-function of Opg on the extracellular matrix of the aortic media.
EPA Reduces JNK Phosphorylation and Mmp-9 Expression in the Media of AAAs. We next assessed JNK phosphorylation and Mmp-9 expression in the AAA model. Consistent with a previous report [15], focal expression of pJNK co-localized with Mmp-9 was observed in Opg-KO mice fed the EPA (-) diet at six weeks (Fig 2A). However, these findings were attenuated in the medial layer of Opg-KO mice fed the EPA (+) diet. Mmp-9 and pJNK double-positive areas were significantly increased in Opg-KO mice fed the EPA (-) diet compared to mice fed the EPA (+) diet (Fig 2B and 2C).
EPA Down-regulates Trail in the Media of AAAs. Trail is up-regulated in the media of AAAs in Opg-KO mice [15], and thus we tested whether Trail expression could be reduced with the EPA (+) diet. In Opg-KO mice fed the EPA (-) diet, Trail was markedly up-regulated in the outer zone of the medial layer, where Mmp-9 was mostly co-localized (Fig 3A). EPA suppressed the up-regulation of Trail and Mmp-9 in the media of Opg-KO mice (Fig 3A and S2 Fig).
We previously reported that Trail could induce its own expression, similar to TNF-α [24], and that its activation could be blocked by the decoy receptor Opg. Given that Trail and p-JNK are expressed in the medial layer, we hypothesized that Trail could induce its own expression through the JNK signaling pathway, and examined whether rh-TRAIL-induced Trail expression is inhibited by JNK inhibitors (AS-601245, SP-600125) in our SMC culture system. Trail expression was reduced in a dose-dependent manner (Fig 3B and 3C), suggesting that Trail induces its own expression via JNK activation in SMCs.
EPA Down-regulates Mmp-9 Expression by Inhibiting the JNK Pathway in SMCs. Upregulation of Mmp-9 expression by rh-TRAIL and TNF-α can occur through the JNK pathway [8], [15]. Here, we tested whether EPA could inhibit the increase in Mmp-9 expression induced by Trail through JNK phosphorylation in SMCs. EPA reduced Trail-induced JNK phosphorylation ( Fig 4A) and Mmp-9 expression (Fig 4B).
EPA has been reported to exert its anti-inflammatory effects by inhibiting the activation of Tak-1, a kinase for JNK and NF-κB, through Gpr-120/Ffar-4 [21]. We first confirmed that Trail activates Tak-1 (Fig 5C). We then tested whether EPA or GW9508, an agonist for Gpr-120/Ffar-4, could reduce the Trail-induced phosphorylation of Tak-1. Tak-1 phosphorylation was completely inhibited with EPA or GW9508 treatment (Fig 5D and 5E). In addition to Tak-1 inactivation, JNK phosphorylation and Mmp-9 expression were inhibited by GW9508 in a dose-dependent manner (Fig 5F and 5G). Thus, stimulation of Gpr-120/Ffar-4 by EPA or GW9508 could inhibit the JNK pathway through inactivation of Tak-1.

Discussion
In this study, we found that EPA prevents disruption of the media in the abdominal aorta of Opg-KO mice. We previously reported that CaCl 2 -induced inflammation was prolonged in Opg-KO mice, and that this resulted in complete destruction of the media [15]. We speculated that, in the absence of Opg, Trail-induced Trail expression forms a vicious cycle that strongly induces Mmp-9 expression. Here, we demonstrated that this vicious cycle is blocked by EPA. Upon investigation of the mechanism underlying this effect, we found that EPA reduces the phosphorylation of Tak-1 and JNK, and down-regulates Mmp-9 expression via Gpr-120/Ffar-4 in cultured SMCs. Epidemiological studies have shown EPA to have inhibitory effects on cardiovascular events, especially in more severe cases [18]. We speculate that, in such cases, the function of OPG might be down-regulated or blocked, and that this insufficiency in OPG activity might worsen the vicious cycle of Trail-induced Trail expression. Our findings suggest that  OPG, a protector of bone metabolism, also plays a protective role in vascular walls, and that EPA can rescue cells from the negative effects of OPG insufficiency.
Recent reports from two groups suggest that EPA plays a preventive role against AAA formation in different animal models. Wang et al. reported that EPA reduced TNF-α-induced MMP-9 expression in macrophages [19], and Yoshihara et al. found that EPA reduced the levels of inflammatory cytokines, such as IL-6 and MCP-1, in macrophages [20]. However, neither group focused on Trail, an inflammatory cytokine, nor did they address the role of SMCs in AAA formation. Our present study, for the first time, demonstrates a potential role for EPA in preventing AAAs and down-regulating Trail in SMCs.
Although EPA blocked severe AAA lesions in Opg-KO mice, it did not affect the formation of AAAs in wild-type mice (Fig 1B). Contrary to our results, Wang et al. showed a significant reduction of CaCl 2 -induced AAAs with EPA treatment. Because CaCl 2 -induced AAA formation is quite drastic, we decided to weaken the effects of CaCl 2 by reducing the amount applied. Specifically, we restricted the contact area by using a small piece of cotton, and limited the treatment duration to 15 minutes. As a result, AAA lesions were weak in wild-type mice, to the extent that no differences were observed between EPA (-) and EPA (+) groups. Wang et al. also used two times higher dose of EPA (10% wt/wt) than that we used in this study. These different experimental conditions might explain the differences between their results and ours.
A mechanism underlying the effects of EPA has been previously proposed. First, activation of Gpr-120/Ffar-4 recruits β-arrestin-2 from the intercellular domain to the plasma membrane, and this is accompanied by the translocation of TAK-1/MAP3K7 binding protein 1 /MAP3-K7IP1 (Tab-1), as a binding partner of β-arrestin-2, in macrophages and adipocytes [22]. Subsequently, Tak-1 and its downstream signaling pathway, including JNK and NF-κB, fails to be activated even in the presence of TNF-α or lipopolysaccharides (LPSs), given the lack of Tab-1 to activate Tak-1. Thus, EPA may inhibit the enlargement of AAAs through activation of Gpr-120/Ffar-4 by inactivating Tak-1 and the downstream JNK pathway in SMCs. This is consistent with a study by Yoshimura et al. that reported the prevention of AAA progression by inhibition of JNK [7].
Based on our results, we propose a model for the inhibitory effects of EPA on AAAs via SMCs (S4 Fig). EPA reduces Trail-induced Mmp-9 expression by inhibiting the Tak-1-JNK pathway through Gpr-120/Ffar-4, thereby preventing the enhancement of AAAs. Our findings may contribute the discovery of therapeutics that suppress the dilatation of aortic aneurysms.  10 ng/mL). Phosphorylation of Tak-1 and JNK from SMCs, which were cultured with rh-TRAIL, was inhibited by EPA (D) and GW9508 (E and F). (G) Relative expression levels of Mmp-9 mRNA in SMCs cultured for six hours with rh-TRAIL (10 ng/mL). Expression of Mmp-9 in SMCs, induced by rh-TRAIL, were inhibited by GW9508. (H-J) Western blotting analysis for pTak-1/Tak-1 (H) and pJNK/JNK (I) from SMCs, which were treated with scrambled (NC) or Gpr-120/Ffar-4 siRNA (10 nM) and then cultured with rh-TRAIL (10 ng/mL) for 15 minutes (H) and 30 minutes (I). (J) Relative expression levels of Mmp-9 mRNA from SMCs treated with scrambled (NC) or Gpr-120/ Ffar-4 siRNA (10 nM) and cultured for six hours with rh-TRAIL (10 ng/mL). Data are presented as mean ± SD; *p<0.05, as compared with controls or Trail alone. Relative expression levels of Gpr-120/Ffar-4 mRNA by qPCR from SMCs cultured with rh-TRAIL (10 mg/mL) for six hours with or without pre-incubation with EPA (10 μM). Gpr-120/Ffar-4 siRNA was applied 24 hours before pre-incubation with EPA. Data are presented as mean ± SD; Ã p<0.05, as compared with control. (EPS)

S4 Fig. Schematic of mechanism by which EPA prevents the development of AAAs.
In Opg-KO mice, inflammation triggered by CaCl 2 induces the up-regulation of inflammatory cytokines, such as Trail, in SMCs of the aorta. Trail activates the Tak-1-JNK pathway through the Trail receptor (Trail-R), inducing the expression of genes encoding proteolytic enzymes for the extracellular matrix, such as Mmp-9 and Trail itself. The enhanced activity of proteolytic enzymes normally allows for remodeling of the aortic walls, leading to aneurysmal dilatation in wild-type mice. On the other hand, with the loss of Opg function, continuous expression of Trail and excessive production of Mmp-9 occur in SMCs. This condition could be a major cause of the disruption of medial elastic fibers. EPA binds to and activates Gpr-120/Ffar-4 as a ligand and inhibits Tak-1, resulting in the inactivation of JNK and reduced expression of Mmp-9. (EPS)