Eicosapentaenoic acid prevents the progression of intracranial aneurysms in rats

In some experiments, secondary antibodies conjugated with horseradish-peroxidase (Vector Laboratories, Burlingame, CA) were used and immune complexes were detected with either a Vectastain Elite ABC kit (Vector Laboratories) or an EnVision kit (DAKO, Glostrup, Denmark) using 3,3′-diaminobenzidine (DAB) as a substrate. Images were acquired using a light microscope (IX71, Olympus) equipped with a UPlanFLN 4×/NA 0.13 objective (Olympus), UPlanFLN 10×/NA 0.3 objective (Olympus), or LUCPlanFLN 40×/NA 0.6 objective (Olympus) at a resolution of 1360 × 1024 pixels with a field of view of 4.4 × 3.3 mm, 1.8 × 1.3 mm, or 438.6 × 330.2 μm, respectively. Images were captured using a digital camera (DP72, Olympus) and a cellSens Standard (version 1.6, Olympus). The brightness and contrast of the images were adjusted in an Adobe Photoshop Elements14 (Adobe Systems).

Primary antibodies used are as follows: rabbit polyclonal anti-GPR120 antibody (dilution 1:900, #LS-A2003, LifeSpan BioSciences, Inc., Seattle, WA), mouse monoclonal anti-CD31 antibody (1:100, #M0823, DAKO), mouse monoclonal anti-smooth muscle α-actin (SMA) antibody (1:200, #M0851, DAKO), and mouse monoclonal anti-CD68 antibody (1:100, #M0814, DAKO).

The secondary antibodies used are as follows: Alexa Fluor 488-conjugated donkey anti-rabbit IgG (H + L) antibody (1:100, #A21206, Invitrogen, Carlsbad, CA) and Alexa Fluor 594-conjugated donkey anti-mouse IgG (H + L) antibody (1:100, #A21203, Invitrogen).

A total of 24 rats were used to examine the pharmacokinetics of eicosapentaenoic acid (EPA). The whole blood was transcardially collected from four rats before EPA administration under intraperitoneal injection of a lethal dose of pentobarbital sodium (200 mg/kg). The remaining 20 rats were randomly divided into the two groups; vehicle-treated group and the EPA-treated group. Ten rats in each group were administrated 1000 mg/kg EPA or olive oil as a vehicle. The whole blood was then collected transcardially at 1 h (n = 3), 6 h (n = 3), or 24 h (n = 4) after the administration.

To induce IA, under general anesthesia by intraperitoneal injection of pentobarbital sodium (50 mg/kg), 7-week-old male rats were subjected to ligation of the left carotid artery and hypervolemia achieved by the combination of a high salt diet and ligation of the left renal artery. Immediately after above surgical manipulations, the chow was completely switched from a normal to the special one containing 8% sodium chloride and 0.12% 3-aminopropionitrile (Tokyo Chemical Industry, Tokyo, Japan), an inhibitor of lysyl oxidase that catalyzes the cross-linking of collagen and elastin. 3-Aminopropionitrile was added in a chow to reduce the stiffness of arterial walls and thus facilitate degenerative changes of the media. The special chow was refilled twice a week, and all rats kept free access to water and the special chow after surgical manipulations during the observation period. A total of 60 rats were subjected to surgical manipulations and 30 rats among them were used for examining the effect of EPA on IAs. Another 30 rats were used to determine the intake of EPA in the plasma. These 30 rats were randomly allocated into three groups: the olive oil-administered group (vehicle-treated group, n = 10), the 100 mg/kg/day EPA-administered group (n = 10), or the 1000 mg/kg/day EPA-administered group (n = 10). Dead animals during the observation period were excluded from the analysis. Blood pressure was measured by the tail-cuff method on the 11th day after surgical manipulations. The rats were placed in a cylindrical warmer (#THC-31, Softron, Tokyo, Japan) and kept at 37 °C during the measurement. Blood pressure were then measured using an automatic measurement device (#BP-98A, Softron) without any anesthesia and its value was defined as the average of three consecutive measurements. Body weight was measured twice; before surgical manipulations and on the day of sacrifice. At times indicated in the corresponding figure legends or results after above surgical manipulations, animals were deeply anesthetized by intraperitoneal injection of pentobarbital sodium (200 mg/kg), and transcardially perfused with 4% paraformaldehyde solution. The bifurcation site of anterior cerebral artery (ACA)-olfactory artery (OA) including the induced IA lesion was then stripped, and serial frozen sections were made. Histopathological examination to evaluate IAs was done after Elastica van Gieson staining which visualizes internal elastic lamina (IEL) and induced IAs were histologically defined as a lesion with the disruption of IEL [8, 13]. The area of IA was measured using ImageJ software (version 1.51s, https://​imagej.​nih.​gov/​ij/​index.​html) [14] using the section including the maximum diameter of the lesions.

EPA (Epadel (ethyl icosapentate), Mochida Pharmaceutical Co., LTD., Tokyo, Japan) was commercially available and we thus purchased this drug. Immediately before administration, EPA was extracted from Epadel capsule and then diluted with olive oil to the concentration of 25 mg/ml or 250 mg/ml. Then, 4 ml/kg of EPA solution was orally administered by gavage. The dose of EPA was determined by referencing the previous reports in a rat model [15‐17]. For the measurement of plasma concentration of EPA, docosahexaenoic acid (DHA), and also arachidonic acid (AA), arterial blood was transcardially collected with an 18-gauge needle at 24 h after the last administration of EPA or vehicle after intraperitoneal injection of a lethal dose of pentobarbital sodium (200 mg/kg). The collected blood (5–8 ml per animal) was heparinized and the plasma was then collected. The plasma samples were stored at − 80 °C until measurement. The plasma concentration of EPA, DHA, and AA was measured by gas chromatography at SRL Inc. (Tokyo, Japan).

The primary antibodies used are as follows: Cy3-conjugated mouse monoclonal anti-smooth muscle α-actin antibody (1:200, #6198-2ML, Sigma Aldrich, St. Louis, MO), mouse monoclonal anti-CD68 antibody (1:100, #31630, Abcam, Cambridge, UK), and rabbit polyclonal anti-C-C motif chemokine 2 (CCL2) antibody (1:100, #ab9779, Abcam).

The secondary antibodies used are as follows: Alexa Fluor 488-conjugated donkey anti-mouse IgG (H + L) antibody (1:100, #A21202, Invitrogen) and Alexa Fluor 488-conjugated donkey anti-rabbit IgG (H + L) antibody (1:100, #A21206, Invitrogen).

Macrophage was defined as the cell positive for CD68 in immunohistochemistry. The number of infiltrated macrophages in each lesion was calculated as a cell count present around the dome of induced aneurysms. Independent two researchers (Y.A. and I.O.) who did not know the group allocation counted the number of CD68-positive cells in each slide and the results were compared. When disagreed, the researchers checked and discussed the image together, and determined the number of positive cells.

Plasma was prepared from rats administered for 12 days with 1000 mg/kg/day of EPA. The plasma was collected at 24 h after the last administration of EPA as described above. And the concentration of Resolvin E1 was measured by a Rat Resolvin E1 ELISA Kit (#MBS2601346, MyBioSource, Inc., San Diego, CA) according to the manufacturer’s instructions.

The Raw264.7 cell line (ATCC, Manassas, VA), a mouse monocyte/macrophage cell line, was used due to endogenous expression of functional GPR120 [11]. Cells were cultured in Dulbecco’s modified Eagle medium (DMEM) (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) supplemented with 10% fetal bovine serum. In some experiments using EPA, DMEM without phenol-red (FUJIFILM Wako Pure Chemical Corporation) was used. Cells were pre-treated with GPR120 agonists, EPA (Mochida Pharmaceutical Co., Ltd.) or TUG-891 (Cayman Chemical, Ann Arbor, MI), before stimulation with lipopolysaccharide (LPS) (Lot. 123M4052V, # L2654, Sigma Aldrich).

Whole cell lysate was prepared by a RIPA buffer (Sigma Aldrich) supplemented with proteinase and phosphatase inhibitors (Roche, Indianapolis, IN). Protein concentration was, then, determined by a bicinchoninic acid (BCA) method (Pierce BCA Protein Assay Kit, Thermo Scientific, Waltham, MA). After sodium dodecyl sulfate-poly-acrylamide gel electrophoresis (SDS-PAGE), separated proteins were transferred to a PDVF membrane (Hybond-P, GE healthcare, Buckinghamshire, UK) and blocked with an ECL plus blocking agent (GE healthcare). The membranes were incubated with primary antibodies followed by incubation with an anti-IgG antibody conjugated with horseradish peroxidase (GE healthcare). Finally, the signal was detected by a chemiluminescent reagent (ECL Prime Western Blotting Detection System, GE healthcare). α-Tubulin was served as an internal control.

Antibodies used in Western blot analysis are as follows: rabbit monoclonal anti-p65 antibody (Cell Signaling Technology, Danvers, MA), rabbit monoclonal anti-phospho NF-κB p65 (Ser536) antibody (Cell Signaling Technology), and mouse monoclonal anti-α-tubulin antibody (Sigma Aldrich).

Primer sets used in the present experiment are as follows: forward 5′-GTCTCTGCCGCCCTTCTGTG-3′ and reverse 5′-AGGTGACTGGGGCATTGATTG-3′ for Ccl2 which encodes monocyte chemoattractant protein 1 (MCP-1), and 5′-ACGACCAGAGGCATACAGGGA-3′ and 5′-CCCTAAGGCCAACCGTGAAA-3′ for Actb.

Data are shown as mean ± SD or by box-and-whisker plots. Differences between the two groups were examined using a non-parametric Mann-Whitney test by a JMP Pro 13 (SAS Institute Inc., Cary, NC). Statistical comparisons between more than two groups were conducted using a Kruskal-Wallis test followed by the post-hoc Steel-Dwass test by a JMP Pro 13 (SAS Institute Inc.). We used above non-parametric tests in the present study because the sample size was not enough to assess whether the data sets were normally distributed or not by a Shapiro-Wilk test. The Jonckheere-Terpstra test was performed to confirm a dose-response relationship using an EZR software (Saitama Medical Center, Jichi Medical University, Saitama, Japan, http://​www.​jichi.​ac.​jp/​saitama-sct/​SaitamaHP.​files/​statmedEN.​html) [18]. A p value smaller than 0.05 was defined as statistically significant.

We picked up GPR120 (free fatty acid receptor 4 (FFAR4)) as a putative therapeutic target from GPCRs whose expression was significantly upregulated in human unruptured IA lesions (n = 4) compared with those in control arterial walls (n = 3) (p = 7.95E⁻¹⁰) from previously obtained comprehensive gene expression profile data by RNA-sequencing analysis [10] (ID number #PRJNA553307 at BioProject database (http://​www.​ncbi.​nlm.​nih.​gov/​bioproject)). To confirm whether over-expression of GPR120 is indeed reflected in protein expression, we performed immunohistochemical analysis using both human IA specimens (n = 4) and control arterial walls (n = 3). In control arterial walls, signals for GPR120 in immunostaining were detected throughout the media and sparsely in the adventitia but not in the intima (Fig. 1 and Additional file 1: fig. S1). In IA lesions, signals for GPR120 in the media became sparse, presumably due to loss of medial smooth muscle cells, but signals in the intima was induced (Fig. 1 and Additional file 1: fig. S1). Positive signals for GPR120 staining were co-localized with the marker of smooth muscle cells, SMA, both in IA lesions and control arterial walls (Fig. 2). In addition, other positive signals for GPR120 staining observed in IA walls were co-localized with the marker of macrophages, CD68, in the adventitia or the marker of endothelial cells, CD31, in the intima (Fig. 2), suggesting the expression of GPR120 in macrophages and the induction of this receptor in endothelial cells during the progression of the disease.

We selected EPA as an agonist for GPR120 [11, 19] because EPA is the already-approved drug for a clinical usage. We first examined the temporal change of the plasma concentration of EPA after the oral administration of this compound to confirm that EPA orally administrated was certainly taken into the plasma in rats. The plasma concentration of EPA before administration was 32.8 ± 2.7 μg/ml (n = 4). The plasma concentration of EPA was certainly and significantly increased by the oral administration of EPA (1000 mg/kg) compared with the concentration in vehicle-treated group (1 h, 38.2 ± 2.0 vs. 139.0 ± 26.1 μg/ml, vehicle-treated group (n = 3) vs. 1000 mg/kg/day-administered group (n = 3), p = 0.049; 6 h, 28.4 ± 1.4 vs. 241.0 ± 66.2 μg/ml, vehicle-treated group (n = 3) vs. 1000 mg/kg/day-administered group (n = 3), p = 0.049; 24 h, 34.7 ± 6.6 vs. 114.8 ± 12.4 μg/ml, vehicle-treated group (n = 4) vs. 1000 mg/kg/day-administered group (n = 4), p = 0.020) (Additional file 1: fig. S2 left). Because DHA is a derivative of EPA [20, 21], the plasma concentration of DHA was also measured to rule out the effect of DHA when EPA was administered. We then found that the plasma concentration of DHA was not increased by the oral administration of EPA (Additional file 1: fig. S2 right). We thereby administered 100 mg/kg/day or 1000 mg/kg/day EPA to the rat model of IAs and again examined the plasma concentration of EPA or related factors on the 5th day after surgical manipulations to induce IAs. We confirmed the increase of EPA (vehicle-treated group, 16.3 [12.3–29.9] μg/ml, n = 9; 100 mg/kg/day-administered group, 32.2 [28.8–34.2] μg/ml, n = 8; 1000 mg/kg/day-administered group, 74.3 [55.6–97.9] μg/ml (median [interquartile range]), n = 9; vehicle-treated group (n = 9) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.0012, 100 mg/kg/day-administered group (n = 8) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.0026) and also the elevation of EPA/AA ratio in a dose-dependent manner (p = 3.80E^-7, the Jonckheere-Terpstra test) (Fig. 3a). The body weight was significantly increased during the observation period in all groups examined (vehicle-treated group, 234.0 ± 5.2 vs. 280.4 ± 16.1 g, pre-treatment vs. post-treatment, n = 9, p = 0.0039; 100 mg/kg/day-administered group, 226.6 ± 6.8 vs. 269.8 ± 9.6 g, n = 10, p =0.002; 1000 mg/kg/day-administered group, 233.8 ± 4.1 vs. 273.3 ± 17.0 g, n = 9, p = 0.0039). Also the administration of EPA did not influence the systolic blood pressure as well (Additional file 1: fig. S3). On the 12th day after IA induction, all rats examined developed IAs characterized by the disruption of an internal elastic lamina in Elastica van Gieson staining (Fig. 3b). Three of 30 rats (10%) surgically manipulated died on the 3rd, 7th, and 11th day after the manipulations, and these animals were excluded from the analyses (vehicle-treated group, n = 2, 1000 mg/kg/day-administered group, n = 1). Thereby, the mortality rate during the observation period was 2/10 (20%) in vehicle-treated group, 0/10 (0%) in 100 mg/kg/day-administered group, and 1/10 (10%) in 1000 mg/kg/day-administered group. Though the incidence of IAs was not influenced by the administration of EPA, the size of the lesions was significantly suppressed in a dose-dependent manner (p = 0.012, the Jonckheere-Terpstra test) (vehicle-treated group, 264.3 [222.1–415.3] μm², n = 8; 100 mg/kg/day-administered group, 247.9 [138.9–349.5] μm², n = 10; 1000 mg/kg/day-administered group, 97.7 [29.5–228.3] μm², n = 9; vehicle-treated group vs. 1000 mg/kg/day-administered group, p = 0.037) (Fig. 3b, c).

We used EPA as an agonist for GPR120. However, as well-known, EPA exerts it action not only via GPR120 as an agonist but also via the formation of the metabolite, Resolvin, with a potent anti-inflammatory effect [22‐27]. We therefore examined the concentration of Resolvin E1 in the plasma collected from EPA-administered animals. The plasma concentration of Resolvin E1 in animals treated with 1000 mg/kg/day of EPA for 12 days was rather significantly lower than that in the vehicle-treated group (2.60 [1.36–4.30] vs. 1.03 [0.56–1.43] ng/ml, vehicle-treated group (n = 8) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.048) (Fig. 3d), supporting that the suppressive effect of EPA on the progression of IAs was via GFP120.

We, next, examined the effect of EPA on degenerative changes of media and inflammatory responses in IA lesions, one of the hallmarks reflecting the disease progression [28, 29]. The thinning of medial smooth muscle cell layer in lesions visualized by immunohistochemistry for a smooth muscle cell marker, SMA, was significantly suppressed by the administration of EPA (thickness of the thinnest portion in lesions; 3.10 [2.23–6.60] vs. 8.83 [5.85–11.7] μm, vehicle-treated group (n = 8) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.0168, 4.23 [3.76–5.67] vs. 8.83 [5.85–11.7] μm, 100 mg/kg/day-administered group (n = 10) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.0134, relative thickness; 0.19 [0.14–0.40] vs. 0.51 [0.41–0.74], vehicle-treated group (n = 8) vs. 1000 mg/kg/day-administered group (n = 9), 0.28 [0.23–0.47] vs. 0.51 [0.41–0.74], 100 mg/kg/day-administered group (n = 10) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.027) (Fig. 4).

We then assessed the infiltration of macrophages which play the crucial role in the pathogenesis of IAs through mediating chronic inflammation [8, 9, 13, 28, 30‐33]. The infiltration of CD68-positive macrophages into IA lesions was significantly suppressed by the administration of EPA (33.2 [22.1–61.0] vs. 0.0 [0.0–22.1] cells/mm², vehicle-treated group (n = 8) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.028, 33.2 [22.1–66.5] vs. 0.0 [0.0–22.1] cells/mm², 100 mg/kg/day-administered group (n = 10) vs. 1000 mg/kg/day-administered group (n = 9), p = 0.013) (Fig. 5).

Referencing the previous report demonstrating endogenous expression of GPR120 in RAW264.7 cell line [11] and considering the crucial role of macrophages in the pathogenesis [9, 30, 31], we used this line to examine the effect of EPA on inflammatory responses observed in in vivo (Fig. 5). We first examined whether EPA indeed could modulate LPS-induced inflammation via GPR120. Referencing the previous study that EPA suppresses activation of NF-κB via GPR120 [11, 34], we examined the effect of exogenous EPA on the phosphorylation of NF-κB p65 subunit at Serine 536 residue by Western blot analysis. Addition of EPA inhibited LPS-induced phosphorylation of NF-κB p65 subunit (Ser536) as expected (Fig. 6a and Additional file 1: fig. S4). We next assessed Ccl2 expression in these cells by RT-PCR analysis considering the pivotal role of macrophages in the pathogenesis [30, 33, 35] and also the inhibition of macrophage infiltration in lesions by EPA (Fig. 5). EPA partially but significantly suppressed LPS-induced Ccl2 expression consistently with the result in NF-κB activation (Fig. 6b). Importantly, a selective GPR120 agonist, TUG-891 [36, 37], reproduced the effect of EPA on LPS-induced Ccl2 expression (Fig. 6b). Since induction of MCP-1 encoded by Ccl2 leads to the formation of a self-amplifying loop among macrophages to exacerbate inflammatory responses by recruiting more macrophages to the inflammatory microenvironment, which in turn promotes the progression of IAs [9], we next examined whether EPA indeed inhibits MCP-1 expression in vivo in a rat model. As a result, CCL2 expression was induced in IA lesions especially at the adventitia and its expression was remarkably suppressed in IA lesions from EPA-administered rats (Fig. 6c), supporting the in vivo relevance of the in vitro study (Fig. 6a, b).

The absolute values of each parameter are listed in the Additional files 2, 3, 4 and 5: Table S1-S4.