Contents lists available at ScienceDirect

Prostaglandins and Other Lipid Mediators

journal homepage: www.elsevier.com/locate/prostaglandins

Omega-3 PUFA attenuate mice myocardial infarction injury by emerging aprotective eicosanoid pattern

Xuan Fanga, Wenbin Caib, Qian Chengb, Ding Aib, Xian Wanga, Bruce D. Hammockc, Yi Zhua,b,Xu Zhanga,b,⁎

a Department of Physiology and Pathophysiology, Peking University Health Science Center, Beijing, Chinab Tianjin Key Laboratory of Metabolic Diseases and Department of Physiology, Tianjin Medical University, Tianjin, Chinac Department of Entomology and Comprehensive Cancer Center, University of California Davis, Davis, California, USA

A R T I C L E I N F O

Keywords:Omega-3 PUFAEicosanoidMyocardial infarctionMetabolomics

A B S T R A C T

Omega-3 polyunsaturated fatty acid (n-3 PUFA) supplementation is a recommended preventive approach againstcardiovascular diseases, but its mechanism of protection against myocardial infarction (MI) injury is not fullyunderstood. Eicosanoid metabolomics demonstrated an abnormal eicosanoid profile was in the plasma of micereceiving MI surgery. 19,20-EDP, 17,18-EEQ, 14,15-EET and 9,10-EpOME were decreased, and PGE2 was in-creased by the surgery. N-3 PUFA-rich diets feeding or transgene of Fat-1 shifted the eicosanoid profile to an n-3PUFA dominant style and attenuated the myocardial infarction injury. Multiple logistic regression analysissuggested the degree of MI injury was related with an eicosanoid pattern, composed by eicosanoids derived fromboth n-3 and n-6 PUFA in the three enzymatic pathways. These results suggested the benefits of n-3 PUFA on MIwas achieved synergistically.

1. Introduction

Myocardial infarction (MI) is an acute and severe heart condition,which is caused by the sudden interruption of the blood circulation ofthe coronary artery and the subsequent damage of myocardium due tothe ischemic status. Although some interventions applied in preventionand therapy were proved effective, MI is still a leading cause of mor-bidity and mortality worldwide [1]. Increasing the dietary intake ofomega-3 polyunsaturated fatty acids (n-3 PUFAs) is a prevalent pre-ventive approach to cardiovascular diseases due to their protective ef-fects [2]. 1 g/day of the two n-3 PUFAs eicosapentaenoic acid (EPA)and docosahexaenoic acid (DHA) is recommended by the AmericanHeart Association/American College of Cardiology, the European So-ciety for Cardiology, et al. for treatment post-myocardial infarction andprevention of sudden cardiac death [3,4]. N-3 PUFAs can easily entercells via fatty acid transporters, rapidly converted to fatty acid acyl-CoAthioesters and incorporated into cell membranes as glyceropho-spholipids (prevalently in the sn-2 position) [5]. The PUFAs, includingn-3 and n-6 (mainly arachidonic acid, AA), in membranes, could bereleased by phospholipase A2 (PLA2) when the cell was activated by avariety of stimuli. Released PUFAs are metabolized by three majorpathways initiated by cyclooxygenases (COX), lipoxygenases (LOX) and

cytochrome P450 (CYP) and produce hundreds of bioactive metabo-lites, which are called eicosanoids [6]. Prostaglandins (PGs), throm-boxanes (TXs), leukotrienes (LTs), lipoxins (LXs), and other epoxy-,hydroxyl- and keto-metabolites are the main components of eicosa-noids. N-3 and n-6 PUFAs compete for the enzymes generating eico-sanoids because of the similarity in their structure. As a result, theuptake of n-3 PUFA shift the profile of eicosanoids and comprehensivelyinfluence their function. In general, 3-series PGs and TXs, 5-series LTs(derived from EPA) have weaker pro-inflammatory, platelet-ag-gregating and vasoconstricting activities compared with 2-series PGsand TXs, 4-series LTs (derived from AA) [7], but the epoxides derivedfrom n-3 PUFAs (EEQs and EDPs) have stronger cardiovascular pro-tective effects than their analogs derived from AA [8]. Also, somespecialized pro-resolving mediators (SPM, including resolvins, pro-tectins and maresins) with potent anti-inflammatory effects could besynthesized from EPA and DHA during the resolution phase of in-flammation [9]. These mechanisms can explain the versatile long-actingprotective effects of n-3 PUFA on cardiovascular.

MI involves many biological processes, such as cardiomyocyteapoptosis and necrosis, myocardial fibrosis, inflammation and so on.Eicosanoids are highly related to these processes, but the eicosanoidprofile in myocardial infarction was not fully understood. Furthermore,

https://doi.org/10.1016/j.prostaglandins.2018.09.002Received 21 April 2018; Received in revised form 17 August 2018; Accepted 5 September 2018

⁎ Corresponding author at: Department of Physiology and Pathophysiology, School of Basic Medical Sciences, Peking University Health Science Center, Beijing100191, China.

E-mail address: xuzhang@tmu.edu.cn (X. Zhang).

Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

Available online 14 September 20181098-8823/ © 2018 Elsevier Inc. All rights reserved.

T

how does n-3 PUFA influence the eicosanoid profile when it is appliedas an intervention approach to MI is entirely unknown. In this work, weused the targeted metabolomic method to study the plasma eicosanoidprofile which was altered by MI surgery and shifted by n-3 PUFA in-tervention. Generally, the protective epoxides, derived from both n-3and n-6 PUFA, were decreased in plasma of the mice receiving MIsurgery. In contrast, PGE2 was increased by MI surgery. These altera-tions were partially reversed by n-3 PUFA-rich diets feeding or trans-gene of Fat-1, which encode the enzyme converting n-6 PUFA to n-3PUFA. It was suggested the benefits of n-3 PUFA was achieved sy-nergistically by affecting the eicosanoid metabolic status. A multiplelogistic regression model was constructed to depict the relationshipbetween the injury degree of MI and the combination of 7 critical ei-cosanoids, including, PGE2, 14,15-EET, 17,18EEQ, 9,10-EpOME, 19,20-EDP, 15-HETE, 17-HDoHE. The metabolic pattern constituted by theseeicosanoids would be an explanation of the protective effect of n-3PUFAs on MI.

2. Materials and methods

2.1. Animal experiments

All procedures of animal handling were approved by theInstitutional Animal Care and Use Committee of Peking UniversityHealth Science Center. Fat-1 tg mice were provided by Dr Alan Zhaofrom Guangdong University of Technology [10]. All mice were main-tained under a 12-h light/dark cycle and controlled temperature withfree access to water. Mice were fed a standard diet [MD12016 (Med-icience, Jiangsu, China)] or fish oil (FO) spiked diet (MD12016 sup-plemented with fish oil, containing 0.5% EPA and 0.35% DHA, w/w)for three weeks. The composition of fatty acids in the FO diet was de-tected by GC–MS and listed in Supplementary Table S1. Mouse MImodel was established with male mice at the age of 6–8 week as pre-viously described [11]. Briefly, C57BL/6 or Fat-1 transgenic mice wereanesthetized with intraperitoneal injection of pentobarbital sodium(60mg/kg). The fourth intercostal space over the left chest was ex-posed, and the heart was rapidly squeezed out of the thoracic space, theleft anterior descending (LAD) artery which below the tip of the leftauricle was tied with a 6 sterile silk suture. The sham subjects under-went the same operation except that the LAD was not ligated. After theoperation, echocardiography was performed with a Vevo 710 RMV-707B (VisualSonics, Toronto, Ontario, Canada)

2.2. TTC staining

Mice were killed with anesthetic seven days after MI operation.Hearts were frozen at − 20 °C for 20min and then were sectioned into5mm thick slices. Five continuous slices from apex to the occlusion sitewere incubated with triphenyl tetrazolium chloride (TTC) at 37 °C for5min. After fixation with 4% paraformaldehyde overnight, each slicewas weighed (w) and photographed with a digital camera. Infarct areas(IR) were indicated as the area not stained by TTC. The IR and leftventricular size (LV) were evaluated by Photoshop. Percentage of in-farct area was calculated as IR/LV= (∑IR(per slice)×w(per slice))/(∑LV(per slice)×w(per slice)).

2.3. Quantitative RT-PCR (qPCR)

Real-time PCR with primers was conducted as previously described[12]. Total RNA from mouse liver was isolated by use of RNAiso Plusreagent (Takara Bio, Japan) as instructed and reverse-transcribed byusing the first-strand cDNA synthesis kit (Thermo Scientific, Rockford,IL, USA). Gene expression was normalized to that of β-actin.

2.4. Sample preparation for metabolomics

Plasma was extracted by solid-phase extraction (SPE) using WatersOasis-HLB cartridges. Before extraction, cartridges were washed withmethanol (1mL) and Milli-Q water (1 mL). Samples were spiked with ISmixture (including 6-keto-PGF1α-d4, PGE2-d4, LTB4-d4, 11,12-DHET-d11, 9-HODE-d4, 20-HETE-d6, 5-HETE-d8, 8,9-EET-d11, ARA-d, EPA-d5, and DHA-d5) and loaded onto cartridges. Cartridges were washedwith 1mL of 5% methanol. The aqueous plug was pulled from the SPEcartridges under high vacuum, and SPE cartridges were further driedunder high vacuum for 20min. Analytes were eluted into tubes with1mL of methanol and evaporated to dryness. The dried residue wasdissolved in 100 μL of 30% acetonitrile and filtered by centrifuge tubefilters (nylon membrane, 0.22 μ m) before LC–MS analysis.

2.5. LC–MS method for metabolomics

Metabolomic analysis by LC–MS/MS for eicosanoids was performedas we previously described [13]. BEH C18 column (1.7 μm,50× 2.1mm i.d., Waters, Milford, MA) was used for ultra-performanceliquid chromatographic separation. Solvent A was water and solvent Bwas acetonitrile. The mobile-phase flow rate was 0.25mL/min. Thegradient was 0–3minutes 30% B; 3–20min B to 60%; 20–24min B to80% and maintained for 3min; and 27–29min B reduced to 30% andmaintained for 1min. The column was maintained at 25 °C, and theinjection volume was set to 10 μL. Targeted profiling of n-6 and n-3PUFA metabolites involved use of a 5500 QTRAP hybrid triple quad-rupole linear ion trap mass spectrometer (AB Sciex, Foster City, CA)equipped with a turbo ion spray electrospray ionization source. Ana-lytes were detected by MRM scans in negative mode. The dwell timeused for all MRM experiments was 25ms. The ion source parameterswere CUR=40 psi, GS1=30 psi, GS2=30 psi, IS = − 4500 V,CAD=MEDIUM, and TEMP=500 °C.

2.6. Data analysis

LC–MS raw data processing involved the use of Analyst 1.6 (ABSciex, Foster City, CA). Eicosanoids were quantified with use ofMultiQuant 2.1 (AB Sciex, Foster City, CA). Metaboanalyst 3.0 [14](http://www.metaboanalyst.ca) was used for metabolomic data ana-lysis, interpretation and visualization, including Wilcoxon Mann-Whitney test, partial least squares discriminant analysis (PLS-DA),clustering and presenting as a heatmap. Spearman’s correlation analysisand multiple logistic regression analysis were performed using R(3.0.3). The correlation network was constructed and analyzed with useof Cytoscape 3.5.1 [15].

3. Results

3.1. The influence of MI surgery on eicosanoid profile

To study the metabolic status of eicosanoids in plasma of mice un-dergoing MI, a mice MI surgery model was established. The left anteriordescending (LAD) was ligated. Seven days after MI operation, heartfunction was measured by echocardiography. Compared with Shamgroup mice, left ventricular interior diameter at end-diastole (LVID;d),left ventricular interior diameter at end- systole (LVID;s), left ven-tricular end-diastolic volume (LV vol;d) and left ventricular end-systolicvolume (LV vol;s) increased significantly in MI group. Also, ejectionfraction (EF%) and fractional shortening (FS%), the two main in-dicators of left ventricular (LV) systolic function, decreased by nearly50% with significance in MI surgery group. (Fig. 1A). Eicosanoids inplasma were extracted and analyzed using metabolomics method. Thelevels of the eicosanoids were listed in Table 1. Heatmap showed theeicosanoid profile in plasma was altered by MI surgery (Fig. 1B).Hierarchical clustering trees beside the heatmap indicated the samples

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

2

in Sham group and MI group could be clustered into two clusters per-fectly by the top 25 changed eicosanoids. PLS-DA was used to develop aclassification model which categorize the 2 type of samples based on MIand Sham surgery. The 2-dimensional score plot of PLS-DA proved theclassification model could perfectly separate the samples by theirgroups (Fig. 1C). Variable importance for prediction (VIP) scores werecalculated, and the top 10 eicosanoids mostly contributing to theclassification were shown on the Y-axis (Fig. 1D). Univariate statistical

analysis (Wilcoxon Mann-Whitney Test) for each eicosanoid was alsoperformed. The significance and the magnitude of changes in eicosa-noids between MI and Sham group were visualized by a volcano plot(Fig. 1E), and the level of each significantly changed eicosanoid in thetwo groups was shown as boxplot (Fig. 1F). Four cardiovascular pro-tective epoxides, both derived from n-3 and n-6, decreased in the MIgroup. PGE2, the COX metabolite of AA, significantly increased in theMI group. Likely all the epoxides are influenced by a common factor, so

Fig. 1. Myocardial infarction surgery led to an abnormal eicosanoid profile in mice plasma.(A) Parameters determined by echocardiography reflecting the decreased heart function after MI surgery. LV, left ventricle; LVID, left ventricle internal diameter;Vol, volume; FS, fraction shortening; EF, ejection fraction; d, diastole; s, systole. *P < 0.05 versus Sham (B) Heatmap showing the altered eicosanoid profile after MIsurgery. Class 1 and 2 represented the Sham and MI group respectively. (C) PLS-DA score plot and (D) features (variables) of top 10 most significant metabolitesbased on VIP scores from loading 1 of PLS-DA. Color bars showed the relative intensities of variables in respective groups. (E) Volcano plot screening out eicosanoidswith significant change in level (x-fold change>2 or<50%, *P < 0.05 versus Sham) in plasma. (F) Boxplot showing the five significantly changed eicosanoids byMI.

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

3

Table 1The levels of the eicosanoids in the plasma of mice (ng/mL, mean ± SEM).

Eicosanoid Sham MIa Sham+FO MI+FOb FAT-1tg Sham FAT-1tg MIc

Epoxide5,6-EET 0.04 ± 0.04 0.02 ± 0.04 N.D. 0.01 ± 0.01 N.D. N.D.8,9-EET 0.07 ± 0.05 0.04 ± 0.03 0.02 ± 0.02 0.01 ± 0.01 0.03 ± 0.01 0.02 ± 0.0111,12-EET 0.36 ± 0.25 0.09 ± 0.03 N.D. 0.01 ± 0.01# 0.04 ± 0.01 0.02 ± 0.0114,15-EET 1.51 ± 0.67 0.32 ± 0.1** 0.24 ± 0.07 0.27 ± 0.12 0.23 ± 0.03 0.23 ± 0.039,10 EpOME 7.01 ± 4.97 1.53 ± 0.36 0.52 ± 0.11 0.65 ± 0.22# 1.47 ± 0.2 1.04 ± 0.1312,13 EpOME 44.04 ± 39.11 5.67 ± 1.01 0.85 ± 0.12 1.22 ± 0.48## 5.35 ± 0.83 4.35 ± 0.5911,12-EEQ 0.04 ± 0.07 N.D. N.D. 0.08 ± 0.08 N.D. 0.03 ± 0.0214,15-EEQ 0.3 ± 0.35 N.D. 0.56 ± 0.11 0.38 ± 0.12# 0.04 ± 0.01 0.05 ± 0.01△

17,18-EEQ 4 ± 2.08 1.06 ± 0.27* 18.72 ± 3.97 10.99 ± 3.99## 1.69 ± 0.21 1.29 ± 0.137,8-EDP 0.43 ± 0.12 0.14 ± 0.07* 0.25 ± 0.11 0.28 ± 0.12 0.28 ± 0.06 0.11 ± 0.0216,17-EDP 1.36 ± 0.38 1.01 ± 0.43 7.4 ± 1.56 5.34 ± 2.21# 1 ± 0.19 0.61 ± 0.1119,20-EDP 4.57 ± 0.74 2.24 ± 0.45** 22.24 ± 4.94 16.52 ± 4.95## 3.48 ± 0.5 2.1 ± 0.3

Diol5,6 DHET 0.18 ± 0.03 0.23 ± 0.11 0.12 ± 0.02 0.09 ± 0.01# 0.14 ± 0.02 0.12 ± 0.028,9 DHET 0.5 ± 0.11 0.52 ± 0.13 0.22 ± 0.02 0.16 ± 0.02## 0.38 ± 0.04 0.3 ± 0.0211,12 DHET 0.89 ± 0.26 0.81 ± 0.21 0.36 ± 0.06 0.26 ± 0.05# 0.33 ± 0.07 0.24 ± 0.03△

14,15 DHET 2.41 ± 0.81 2.64 ± 0.72 1.05 ± 0.15 0.78 ± 0.19# 0.89 ± 0.2 0.59 ± 0.08△

9,10 diHOME 14.97 ± 5.23 14.19 ± 2 2.21 ± 0.35 1.84 ± 0.59## 7.42 ± 1.48 3.94 ± 0.5△

12,13 diHOME 13.45 ± 6.91 10.28 ± 2.96 1.69 ± 0.41 1.4 ± 0.42## 9.8 ± 2.24 4.16 ± 0.58△

5,6-DiHETE 0.22 ± 0.1 0.22 ± 0.12 1.53 ± 0.47 1.06 ± 0.09## 0.47 ± 0.03 0.49 ± 0.04△

14,15-DiHETE 0.49 ± 0.16 0.55 ± 0.21 3.85 ± 0.78 2.25 ± 0.29## 0.74 ± 0.14 0.49 ± 0.0317,18-DiHETE 1.53 ± 0.37 2.44 ± 1.18 8.98 ± 2.61 5.8 ± 0.81# 2.29 ± 0.29 2.1 ± 0.14

Hydroxyl-5-HETE 3.71 ± 0.77 3.04 ± 0.32 3.41 ± 0.53 2.57 ± 0.48 1.97 ± 0.56 1.24 ± 0.17△

8-HETE 1.85 ± 0.42 1.77 ± 0.61 1.62 ± 0.47 0.81 ± 0.14# 0.91 ± 0.1 0.74 ± 0.09△

9-HETE 2.59 ± 0.81 2.46 ± 2.68 5.84 ± 4.44 0.21 ± 0.16 0.7 ± 0.26 0.82 ± 0.3611-HETE 1.5 ± 0.51 1.43 ± 0.64 1.19 ± 0.24 0.77 ± 0.16 0.81 ± 0.16 0.57 ± 0.1512-HETE 36.32 ± 10.31 39.94 ± 39.57 95.6 ± 70.4 11.64 ± 3.29 9.02 ± 1.85 13.01 ± 2.8515-HETE 3.93 ± 0.93 3.14 ± 1.02 3.15 ± 0.45 2.41 ± 0.54 1.69 ± 0.3 1.19 ± 0.23△

16-HETE 0.13 ± 0.02 0.08 ± 0.04 0.03 ± 0.03 N.D.# 0.05 ± 0.02 0.05 ± 0.0117-HETE 0.31 ± 0.04 0.27 ± 0.04 0.28 ± 0.04 0.18 ± 0.03 1.04 ± 0.18 0.72 ± 0.09△

18-HETE 1.07 ± 0.26 0.95 ± 0.1 2.09 ± 0.37 1.26 ± 0.22 0.48 ± 0.11 0.28 ± 0.02△

19-HETE 1.6 ± 0.47 1.14 ± 0.37 0.43 ± 0.12 0.14 ± 0.1# 0.33 ± 0.12 0.16 ± 0.07△

20-HETE 0.33 ± 0.05 0.3 ± 0.1 0.28 ± 0.08 0.18 ± 0.05 0.15 ± 0.05 0.15 ± 0.01△

9-HODE 2.11 ± 0.43 2.48 ± 0.43 1.62 ± 0.64 0.73 ± 0.09## 1.61 ± 0.19 1.22 ± 0.14△

13-HODE 1.66 ± 0.36 1.95 ± 0.31 1.3 ± 0.54 0.56 ± 0.05## 1.11 ± 0.14 0.85 ± 0.1△

5-HEPE 2.39 ± 0.68 1.75 ± 0.44 26.25 ± 6.85 11.74 ± 2.38## 1.37 ± 0.16 1.23 ± 0.138-HEPE 0.98 ± 0.18 0.72 ± 0.08 6.3 ± 1.59 2.89 ± 0.52## 0.74 ± 0.09 0.63 ± 0.079-HEPE 0.74 ± 0.16 0.47 ± 0.04 10.15 ± 3.58 3.09 ± 0.74## 0.68 ± 0.14 0.62 ± 0.0811-HEPE 0.61 ± 0.19 0.61 ± 0.07 6.4 ± 1.09 3.69 ± 0.82## 0.6 ± 0.1 0.52 ± 0.0512-HEPE 5.85 ± 1.6 3.29 ± 2.41 191.09 ± 82.71 39.73 ± 13.45# 7.05 ± 2.01 6.63 ± 1.1315-HEPE 1.78 ± 0.54 1.45 ± 0.26 14.9 ± 2.1 6.5 ± 1.46## 1.3 ± 0.22 0.94 ± 0.1318-HEPE 0.46 ± 0.1 0.56 ± 0.13 5.18 ± 0.78 3.96 ± 1.14## 0.97 ± 0.11 0.69 ± 0.024-HDoHE 2.43 ± 0.58 1.83 ± 0.37 8.35 ± 2.69 10.83 ± 3.82## 2.73 ± 0.42 1.84 ± 0.197-HDoHE 0.47 ± 0.13 0.5 ± 0.15 1.83 ± 0.26 1.44 ± 0.34# 0.5 ± 0.06 0.39 ± 0.058-HDoHE 3.3 ± 1 2.24 ± 1.91 27.38 ± 9.08 9.69 ± 3.38# 2.11 ± 0.59 1.82 ± 0.3510-HDoHE 2.56 ± 1 3.83 ± 2.37 16.83 ± 4.17 8.54 ± 2.96 1.37 ± 0.18 1.29 ± 0.2411-HDoHE 1.55 ± 0.36 1.23 ± 0.15 6.13 ± 1.17 5.37 ± 1.46## 1.24 ± 0.19 0.9 ± 0.0913-HDoHE 0.9 ± 0.49 0.93 ± 0.39 6.02 ± 0.95 4.24 ± 1.15# 0.98 ± 0.19 0.63 ± 0.1314-HDoHE 9.71 ± 2.91 6.58 ± 5.46 82.87 ± 27.53 26.55 ± 9.64# 6.5 ± 1.99 5.46 ± 0.9916-HDoHE 3.57 ± 0.89 2.32 ± 0.36 14.64 ± 2.38 12.18 ± 3.13## 2.27 ± 0.38 1.42 ± 0.2517-HDoHE 4.75 ± 1.08 3.43 ± 1.14 29.66 ± 7.21 16.05 ± 4.44# 3.44 ± 0.55 2.36 ± 0.3220-HDoHE 5.1 ± 1.39 3.15 ± 0.49 24.75 ± 4.78 19.01 ± 4.76## 4.59 ± 0.68 2.97 ± 0.48

PG and others6k-PGF1a 0.23 ± 0.08 0.24 ± 0.14 0.13 ± 0.03 0.09 ± 0.01 0.18 ± 0.06 0.09 ± 0.01PGB2 N.D. N.D. N.D. N.D. 0.02 ± 0.02 N.D.PGD2 0.06 ± 0.11 0.46 ± 0.47 0.07 ± 0.07 0.04 ± 0.04 N.D. N.D.PGE2 0.04 ± 0.07 0.34 ± 0.3* 0.07 ± 0.07 0.03 ± 0.03 0.02 ± 0.02 N.D. △

PGJ2 0.01 ± 0.01 0.02 ± 0.02 N.D. N.D. 0.03 ± 0.01 N.D.15d-PGJ2 0.96 ± 0.18 0.9 ± 0.29 0.26 ± 0.04 0.27 ± 0.12# 0.68 ± 0.09 0.54 ± 0.09TXB2 0.58 ± 0.16 1.27 ± 1.04 0.63 ± 0.38 0.51 ± 0.25 0.48 ± 0.14 0.31 ± 0.08TXB3 N.D. N.D. 0.04 ± 0.01 0.03 ± 0.02 N.D. N.D.5-oxo-ETE N.D. 0.02 ± 0.03 0.09 ± 0.06 0.26 ± 0.2 0.12 ± 0.04 N.D.15-oxo-ETE 0.32 ± 0.08 0.2 ± 0.06 0.22 ± 0.09 0.63 ± 0.34 0.28 ± 0.09 0.1 ± 0.016R-LXA4 0.14 ± 0.1 0.05 ± 0.03 1.17 ± 0.69 1.35 ± 0.82# 0.55 ± 0.08 0.22 ± 0.0910,17-DiHDoHE N.D. N.D. 1.5 ± 0.44 0.95 ± 0.48 N.D. N.D.

a Significance level for the comparison between Sham and MI using Wilcoxon Mann-Whitney test, * p < 0.05, ** p < 0.01.b Significance level for the comparison between MI and MI+ FO using Wilcoxon Mann-Whitney test, # p < 0.05, ## p < 0.01.c Significance level for the comparison between MI and FAT-1tg MI using Wilcoxon Mann-Whitney test, △ p < 0.05.

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

4

Fig. 2. N-3 PUFA supplementation protected mice against myocardial infarction injury and shifted the eicosanoid profile.(A) Parameters determined by echocardiography reflecting the reversed heat function by n-3 PUFA supplementation. *P < 0.05 versus Sham, #P < 0.05 versus MI(B) TTC staining showing the diminished infarction area by n-3 PUFA supplementation. *P < 0.05 versus Sham, #P < 0.05 versus MI. (C) Heatmap showing theshifted eicosanoid profile by n-3 PUFA supplementation. Class 1, Sham group fed with normal diet; Class 2, MI group fed with normal diet; Class 3, Sham group fedwith high n-3 PUFA diet; Class 4, MI group fed with high n-3 PUFA diet. (D) 3D PLS-DA score plot and (E) features (variables) of top 10 most significant metabolitesbased on VIP scores from loading 1 of PLS-DA. Color bars showed the relative intensities of variables in respective groups. (F) Boxplot showing the level of the 5 MIaltered eicosanoids changed by n-3 PUFA supplementation. *P < 0.05 versus Sham, #P < 0.05 versus MI

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

5

we tested transcription level of the epoxide related genes in mice liver,where the epoxides mainly produced. qPCR experiment indicated thedecrease of the four epoxides could be caused by the down-regulation ofCYP2C40 transcription (Fig. S1).

3.2. N-3 PUFA supplementation protected mice from MI injury and shiftedthe eicosanoid profile

To determine the effect of n-3 PUFAs on myocardial infarction, wefed mice with an FO diet or a normal diet for three weeks before MIsurgery. The data of “Sham” and “MI” groups presented in Fig. 1 wereused as control groups in this section of the analysis. Echocardiographyrevealed the heart function was greatly ameliorated by n-3 PUFA(Fig. 2A). The image of TTC staining also showed the area of myo-cardial infarction was diminished in the n-3 PUFA-rich group (Fig. 2B).We applied the metabolomic method to study the eicosanoid profile inthe mice fed with FO diet. The levels of the eicosanoids were listed inTable 1. The heatmap showed FO diet shifted the eicosanoid profile toan n-3 PUFA dominant style (Fig. 2C). Hierarchical clustering analysisshowed the eicosanoid profiles of mice fed with FO mice and normaldiet were in two separated clusters. PLS-DA was also performed withthe eicosanoid profile of the 4 groups (Sham, MI, Sham+FO andMI+FO). The 3- dimensional score plot of PLS-DA showed the micefed with FO or normal diet could be separated by Component 1, andmice receiving MI surgery or not in each diet fed group could be furtherseparated by Component 2 (Fig. 2D). It meant FO supplementationshifted the alteration on eicosanoid profile induced by MI surgery. VIPscores showed the top 10 eicosanoids contributing to the classificationbased on Component 1(Fig. 2E). Univariate statistical analysis for eacheicosanoid was performed, and the boxplots of the five eicosanoidssignificantly changed by MI surgery were shown as Fig. 2F. WilcoxonMann-Whitney Test performed between the MI and MI+ FO groupindicated the level of 19,20-EDP and 17,18-EEQ, the two n-3 PUFAderived eicosanoid, were significantly reversed by FO diet and evenhigher than the Sham group. For n-6 PUFA derived epoxides, the levelof 9,10-EpOME were decreased by FO diet, but the proportion of de-crease in these two eicosanoids is less than the proportion of the in-crease in 19,20-EDP and 17,18-EEQ. The levels of 14,15-EET(p=0.164) and PGE2 (p=0.063) did not reach the significance.

3.3. Fat-1 transgenic mice have a resistance on MI injury and an n-3 PUFAdominant eicosanoid profile

To further validate the effect of n-3 PUFA on myocardial infarction.Fat-1 transgenic mice expressing the C. elegans Fat-1 gene received theMI surgery. The Fat-1 gene encodes an n-3 fatty acid desaturase whichconverts n-6 to n-3 fatty acids (which is absent in mammals). The dataof “Sham” and “MI” groups presented in Fig. 1 were used as controlgroups in this section of the analysis. Echocardiography and TTCstaining showed Fat-1 transgenic mice had a stronger heart functionand a smaller infarction area compared with WT mice receiving the MIsurgery (Fig. 3A and 3B). Eicosanoid metabolomics also demonstratedFat-1 mice had an intrinsic n-3 PUFA dominant eicosanoid profile(Fig. 3C and Table 1). PLS-DA analysis indicated the increase of n-3PUFA derived eicosanoids were still the most important factor con-tributing to the difference of eicosanoid profile (Fig. 3D and E). Theboxplots of the five eicosanoids significantly changed by MI surgerywere shown in Fig. 3F. Wilcoxon Mann-Whitney Test performed be-tween the MI and Fat-1 MI group indicated only PGE2 was significantlyreversed in the Fat-1 MI group, and all the epoxides did not reach thestatistical significance. These results suggested both n-3 PUFA supple-mentation and Fat-1 transgene had a protective effect on MI and pro-duced an n-3 PUFA dominant eicosanoid profile, but the specific sig-nificantly changed eicosanoids between these two n-3 PUFAintervention approaches were different (Table 1).

3.4. Correlation between eicosanoids and the protective eicosanoid patternon MI

Firstly, we tested if the levels of PUFAs in plasma were changed byMI and n-3 PUFA supplementation or Fat-1 transgene. As shown in theboxplot (Fig. 4A), n-3 PUFA supplementation significantly increase thelevel of EPA and DHA and decrease the level of AA. However, Fat-1transgene only increased the level of EPA significantly, but the changeof the DHA and AA level did not reach the significance statistic level.Compared the MI group with the Sham group, the change of PUFAs didnot reach the significance statistic level either. These results couldn’tprove if the protective effect of the two n-3 PUFA intervention ap-proaches on MI was related to the level change of the PUFAs. Accordingto the metabolomics results, we proved both supplementation and Fat-1transgene shifted the eicosanoid profile (Fig. 2C and 3C). However, thefive significantly changed eicosanoid by MI were not reversed by thetwo intervention approaches consistently. We still couldn’t identify thespecific eicosanoids contributed to the protective effect on MI. Con-sidering the eicosanoids are derived from a same metabolic network,we wanted to study the interaction between eicosanoids and the posi-tion of the five significantly changed eicosanoids in the metabolicnetwork. We also wanted to know which parts of the metabolic networkwere influenced by the two intervention approaches and found the ei-cosanoid pattern which contributes to the protective effect. We calcu-lated the Spearman’s correlation coefficient for each correlation be-tween eicosanoids based on the amount of whom in the six groups. Anetwork of eicosanoid was constructed to present the overall correla-tions (Fig. 4B), from which we found most n-3 PUFA derived eicosa-noids were negatively correlated with the n-6 PUFA derived ones(shown by blue edge) and eicosanoids derived from the same PUFAwere positively correlated with each other (shown by black edge). Theeicosanoids influenced by either intervention approach were with redborders. To understand the protective metabolic state, we dissected theeicosanoid network. The network was constituted by five modules,which contain LOX and CYP metabolites derived from n-3 PUFA andCOX, LOX and CYP metabolites derived from n-6 PUFA. The eicosa-noids in the same module positively correlated with each other. Thefive significantly altered eicosanoids (PGE2, 19,20-EDP, 17,18-EEQ,14,15-EET and 9,10-EpOME, shown by yellow nodes) by MI surgerybelonged to 3 modules including CYP n-3, CYP n-6 and COX n-6. Thetwo n-3 PUFA interventions not only reversed four eicosanoids sig-nificantly changed by MI but also changed lots of eicosanoids in all thefive modules. We could not ascertain if the protective effect n-3 PUFAcomes from the reverse of the five eicosanoids of from the regulation ofother eicosanoids.

Although the specific function of eicosanoids in the process ofomega-3 PUFA’s protection on MI could not be identified, we tried tofind which eicosanoids could be the effective factors on cardiac func-tion. Whether the ejection fraction (EF) exceeds 50% was a criterion ofcardiac dysfunction. Logistic regression analysis was used to measurethe influence of each eicosanoid on cardiac function. The area undercurve (AUC) of the receiver operating characteristic (ROC) for eacheicosanoid was listed in Table S2. Most of the AUCs for eicosanoidswere below 0.7. Furthermore, we tested the combinational effect of thefive significantly changed eicosanoids by MI on cardiac function using amultiple logistic regression model (Model 1), in which each eicosanoidwas an independent variable and whether cardiac dysfunction occurswas the dependent variable. The ROC curve suggested this model havegood efficacy in the classification of cardiac function, of which the AUCwas 0.817 (Fig. 4C). Considering the correlation among eicosanoids, wefurther inspected the efficacy by other combinations of eicosanoids. 15-HETE and 17-HDoHE had the highest connection degree in the LOX n-6and LOX n-3 modules respectively, so these two representative eicosa-noids were added to the logistic model (updated as Model 2). Thus,eicosanoids in Model 2 covered all the five modules in the eicosanoidnetwork. The ROC curve of Model 2 indicated the efficacy of

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

6

classification was increased when the 2 LOX metabolites were included(Fig. 4D). The increased efficacy on the classification of cardiac func-tion by multiple logistic regression indicated the factor contributing tothe protective effect of omega-3 PUFA on MI was the eicosanoid patternbut not a specific eicosanoid.

4. Discussion

Although much work reported the cardiovascular protective effectof n-3 PUFA, no studies focused on the perspective from eicosanoidprofile which contains the main metabolite in the three major enzy-matic pathways. Targeted metabolomics method provided

Fig. 3. Fat-1 transgenic mice had resistance on myocardial infarction and showed a n-3 dominant eicosanoid profile style.(A) Parameters determined by echocardiography reflecting the reversed heat function by the Fat-1 transgene. *P < 0.05 versus Sham, #P < 0.05 versus MI. (B) TTCstaining showing the diminished infarction area by the Fat-1 transgene. *P < 0.05 versus Sham, #P < 0.05 versus MI. (C) Heatmap showing the n-3 dominanteicosanoid profile by the Fat-1 transgene. Class 1, Sham group fed with normal diet; Class 2, MI group fed with normal diet; Class 5, Sham group fed with high n-3PUFA diet; Class 6, MI group fed with high n-3 PUFA diet. (D) 3D PLS-DA score plot and (E) features (variables) of top 10 most significant metabolites based on VIPscores from loading 1 of PLS-DA. Color bars showed the relative intensities of variables in respective groups. (F) Boxplot showing the level of the 5 MI alteredeicosanoids changed by the Fat-1 transgene. *P < 0.05 versus Sham, #P < 0.05 versus MI

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

7

comprehensive information about eicosanoids in plasma of mice fedwith n-3 PUFA-rich diet and Fat-1 transgene. The significant eicosa-noids changed by these two kinds of n-3 intervention approach werenot the same. Most of the significant eicosanoids changed by Fat-1transgene were the decreased n-6 PUFA derived ones, whereas n-3PUFA supplementation not only decreased n-6 PUFA derived eicosa-noids, but also increased lots of n-3 PUFA derived eicosanoids (Table 1).Also, the fold-changes of n-3 PUFA derived eicosanoids increased by n-3PUFA supplementation were higher than that increased by the Fat-1transgene. One possible reason was that n-3 PUFA intake from foodmade the level of n-3 PUFA in plasma rise sharply, which lead to a lot ofn-3 PUFA entered tissues and cells and be rapidly metabolized intoeicosanoids by the corresponding enzymes. Whereas Fat-1 transgeneconverses n-6 PUFA to n-3 PUFA in a mild manner, the increase of then-3 PUFA derived eicosanoids is less obvious. The level of EPA andDHA, which was not remarkably changed by the Fat-1 transgene, alsosupported this deduction (Fig. 4A).

Many eicosanoids changed by n-3 PUFA intervention have a definitefunction on the cardiovascular system. For example, EEQs and EDPswere increased by n-3 PUFA intervention. Besides the anti-in-flammatory effect which was broadly reported, both 17,18-EEQ and19,20-EDP were reported to inhibit the calcium-induced arrhythmias incardiomyocytes [16,17]. PGE2 was increased after seven days of MIsurgery and decreased by n-3 PUFA intervention. In most instances,PGE2 was recognized as an inflammation mediator [18]. Deletion ofmicrosomal prostaglandin E2 synthase-1 (mPGES-1) reduced plaqueburden in fat-fed LDLR(-/-) mice, which suggested PGE2 would be apro-atherogenic factor [19]. However, some work reported the poten-tial protective effect of PGE2 on myocardial infarction injury. Deletionof mPGES-1 leads to eccentric cardiac myocyte hypertrophy, LV dila-tion, and impaired LV contractile function after acute MI [20].

Therefore, the actual effects of PGE2 on prevention and repair phases ofmyocardial infarction need to be comprehensively considered. N-3supplementation also elevated most of the monohydroxy- eicosanoid,despite the functional studies of these compounds in the heart is lack.Bone marrow transplantation experiments revealed 18-HEPE from Fat-1 transgenic macrophages exhibited resistance to pressure overload-induced inflammation and fibrosis and ameliorated cardiac function[21]. N-3 PUFA derived specialized pro-resolving mediators (SPM)were not detected in this work, which could be related to the time pointwe selected. SPMs are usually produced in the resolution phase [22].

After acquiring the eicosanoid profile, logistic regression analysiswas applied to depict the relation of eicosanoids and the recovery ofcardiac function after MI surgery. The ROC curve of the logistic modelsindicated the classification efficacy on cardiac function using a com-bination of eicosanoids was better than using each eicosanoid or PUFAsthemselves, which suggested n-3 PUFA synergistically exerted its pro-tective effect. This viewpoint is supported by lots of works in whichspecific eicosanoid metabolites were proved to play a role in differentbiological processes related to MI injury or recovery respectively. Theeicosanoids are derived from a same metabolic network. As a result, theincrease of the substrates could influence lots of metabolites’ level. Apattern composed of seven eicosanoids was suggested to be highly re-lated to the cardiac function. The discovery of this pattern could affordsome inspiration for clinical application, which is the influence on ei-cosanoids should be comprehensively considered when using n-3 PUFAintervention. This pattern could also be an explanation of why someclinical trial of n-3 PUFA supplementation on MI acquired negativeresults [23,24]. Although the specific function of eicosanoids in thepattern need to be further demonstrated, it still suggested the regulationof muti-eicosanoids in a combinational manner would be a new per-spective in clinical application.

Fig. 4. The eicosanoid pattern associated with the protective effect of n-3 PUFA against MI.(A) Boxplot showing the level of PUFAs influenced by n-3 PUFA supplementation and Fat-1 transgene. *P < 0.05 versus Sham. (B) Correlation network of eico-sanoids showing the relations between eicosanoids. (Spearman’s correlation coefficient> 0.5 or< 0.5). Black and blue edges, respectively represent positive andnegative correlations. Pink circle, COX metabolites of n-6 PUFA; pink square, LOX metabolites of n-6 PUFA; pink hexagon, CYP450 metabolites of n-6 PUFA; bluesquare, LOX metabolites of n-3 PUFA; blue hexagon, CYP450 metabolites of n-3 PUFA. The eicosanoids significant changed by MI were in yellow. The eicosanoidsinfluenced by either n-3 PUFA intervention approach were with red borders (C) ROC curve showing the performance of the multivariate logistic regression model 1 inpredicting the ejection fraction. (D) ROC curve showing the performance of the multivariate logistic regression model 2 in predicting the ejection fraction.

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

8

Acknowledgments

This work was supported in part by grants from the National NaturalScience Foundation of China [91539108 and 81870188 to X.Z.,81420108003 to Y.Z., 91439206 and 31230035 to X.W.], TianjinMunicipal Science and Technology Project [14JCYBJC41800 to D.A.]and the Ministry of Science and technology of China[2016YFC0903000] to Y.Z.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in theonline version, at https://doi.org/10.1016/j.prostaglandins.2018.09.002.

References

[1] E.J. Benjamin, M.J. Blaha, S.E. Chiuve, M. Cushman, S.R. Das, R. Deo, S.D. deFerranti, J. Floyd, M. Fornage, C. Gillespie, C.R. Isasi, M.C. Jimenez, L.C. Jordan,S.E. Judd, D. Lackland, J.H. Lichtman, L. Lisabeth, S. Liu, C.T. Longenecker,R.H. Mackey, K. Matsushita, D. Mozaffarian, M.E. Mussolino, K. Nasir,R.W. Neumar, L. Palaniappan, D.K. Pandey, R.R. Thiagarajan, M.J. Reeves,M. Ritchey, C.J. Rodriguez, G.A. Roth, W.D. Rosamond, C. Sasson, A. Towfighi,C.W. Tsao, M.B. Turner, S.S. Virani, J.H. Voeks, J.Z. Willey, J.T. Wilkins, J.H. Wu,H.M. Alger, S.S. Wong, P. Muntner, C. American Heart Association Statistics, S.Stroke Statistics, Heart disease and stroke statistics-2017 update: a report from theAmerican heart association, Circulation 135 (10) (2017) e146–e603.

[2] D. Mozaffarian, J.H. Wu, Omega-3 fatty acids and cardiovascular disease: effects onrisk factors, molecular pathways, and clinical events, J. Am. Coll. Cardiol. 58 (20)(2011) 2047–2067.

[3] S.C. Smith Jr., J. Allen, S.N. Blair, R.O. Bonow, L.M. Brass, G.C. Fonarow,S.M. Grundy, L. Hiratzka, D. Jones, H.M. Krumholz, L. Mosca, R.C. Pasternak,T. Pearson, M.A. Pfeffer, K.A. Taubert, L. Aha/Acc, National Heart, I, Blood, AHA/ACC guidelines for secondary prevention for patients with coronary and otheratherosclerotic vascular disease: 2006 update: endorsed by the National Heart,Lung, and Blood Institute, Circulation 113 (19) (2006) 2363–2372.

[4] G. De Backer, E. Ambrosioni, K. Borch-Johnsen, C. Brotons, R. Cifkova,J. Dallongeville, S. Ebrahim, O. Faergeman, I. Graham, G. Mancia, V. Manger Cats,K. Orth-Gomer, J. Perk, K. Pyorala, J.L. Rodicio, S. Sans, V. Sansoy, U. Sechtem,S. Silber, T. Thomsen, D. Wood, E. Third Joint Task Force of, P. Other Societies onCardiovascular Disease Prevention in Clinical, European guidelines on cardiovas-cular disease prevention in clinical practice. Third joint task force of European andother societies on cardiovascular disease prevention in clinical practice, Eur. Heart.J. 24 (17) (2003) 1601–1610.

[5] R.De Caterina, n-3 fatty acids in cardiovascular disease, N. Engl. J. Med. 364 (25)(2011) 2439–2450.

[6] M. Guichardant, C. Calzada, N. Bernoud-Hubac, M. Lagarde, E. Vericel, Omega-3polyunsaturated fatty acids and oxygenated metabolism in atherothrombosis,Biochim. Biophys. Acta 1851 (4) (2015) 485–495.

[7] D. Bagga, L. Wang, R. Farias-Eisner, J.A. Glaspy, S.T. Reddy, Differential effects ofprostaglandin derived from omega-6 and omega-3 polyunsaturated fatty acids onCOX-2 expression and IL-6 secretion, Proc. Natl. Acad. Sci. U. S. A. 100 (4) (2003)1751–1756.

[8] A. Konkel, W.H. Schunck, Role of cytochrome P450 enzymes in the bioactivation ofpolyunsaturated fatty acids, Biochim. Biophys. Acta 1814 (1) (2011) 210–222.

[9] M.C. Basil, B.D. Levy, Specialized pro-resolving mediators: endogenous regulatorsof infection and inflammation, Nat. Rev. Immunol. 16 (1) (2016) 51–67.

[10] D. Wei, J. Li, M. Shen, W. Jia, N. Chen, T. Chen, D. Su, H. Tian, S. Zheng, Y. Dai,A. Zhao, Cellular production of n-3 PUFAs and reduction of n-6-to-n-3 ratios in thepancreatic beta-cells and islets enhance insulin secretion and confer protectionagainst cytokine-induced cell death, Diabetes 59 (2) (2010) 471–478.

[11] E. Gao, Y.H. Lei, X. Shang, Z.M. Huang, L. Zuo, M. Boucher, Q. Fan, J.K. Chuprun,X.L. Ma, W.J. Koch, A novel and efficient model of coronary artery ligation andmyocardial infarction in the mouse, Circ. Res. 107 (12) (2010) 1445–1453.

[12] S.S. Xue, J.L. He, X. Zhang, Y.J. Liu, F.X. Xue, C.J. Wang, D. Ai, Y. Zhu,Metabolomic analysis revealed the role of DNA methylation in the balance of ara-chidonic acid metabolism and endothelial activation, Biochim. Biophys. Acta 1851(10) (2015) 1317–1326.

[13] Q. Bao, Y. Liu, H. Song, N. Yang, D. Ai, Y. Zhu, X. Zhang, Spectrum evaluation-assisted eicosanoid metabolomics for global eicosanoid profiling in human vascularendothelial cells, Clin. Exp. Pharmacol. Physiol. 45 (1) (2018) 98–108.

[14] J. Xia, D.S. Wishart, Using MetaboAnalyst 3.0 for comprehensive metabolomicsdata analysis, Curr. Protocols Bioinform. 55 (2016) 14 10 1-14 10 91.

[15] P. Shannon, A. Markiel, O. Ozier, N.S. Baliga, J.T. Wang, D. Ramage, N. Amin,B. Schwikowski, T. Ideker, Cytoscape: a software environment for integratedmodels of biomolecular interaction networks, Genome Res. 13 (11) (2003)2498–2504.

[16] J.R. Falck, G. Wallukat, N. Puli, M. Goli, C. Arnold, A. Konkel, M. Rothe, R. Fischer,D.N. Muller, W.H. Schunck, 17(R),18(S)-epoxyeicosatetraenoic acid, a potent ei-cosapentaenoic acid (EPA) derived regulator of cardiomyocyte contraction: struc-ture-activity relationships and stable analogues, J. Med. Chem. 54 (12) (2011)4109–4118.

[17] C. Arnold, M. Markovic, K. Blossey, G. Wallukat, R. Fischer, R. Dechend, A. Konkel,C. von Schacky, F.C. Luft, D.N. Muller, M. Rothe, W.H. Schunck, Arachidonic acid-metabolizing cytochrome P450 enzymes are targets of {omega}-3 fatty acids, J.Biol. Chem. 285 (43) (2010) 32720–32733.

[18] I. Gomez, N. Foudi, D. Longrois, X. Norel, The role of prostaglandin E2 in humanvascular inflammation, Prostaglandins Leukot. Essent. Fat. Acids 89 (2-3) (2013)55–63.

[19] M. Wang, A.M. Zukas, Y. Hui, E. Ricciotti, E. Pure, G.A. FitzGerald, Deletion ofmicrosomal prostaglandin E synthase-1 augments prostacyclin and retards ather-ogenesis, Proc. Natl. Acad. Sci. U. S. A. 103 (39) (2006) 14507–14512.

[20] N. Degousee, S. Fazel, D. Angoulvant, E. Stefanski, S.C. Pawelzik, M. Korotkova,S. Arab, P. Liu, T.F. Lindsay, S. Zhuo, J. Butany, R.K. Li, L. Audoly, R. Schmidt,C. Angioni, G. Geisslinger, P.J. Jakobsson, B.B. Rubin, Microsomal prostaglandin E2synthase-1 deletion leads to adverse left ventricular remodeling after myocardialinfarction, Circulation 117 (13) (2008) 1701–1710.

[21] J. Endo, M. Sano, Y. Isobe, K. Fukuda, J.X. Kang, H. Arai, M. Arita, 18-HEPE, an n-3fatty acid metabolite released by macrophages, prevents pressure overload-inducedmaladaptive cardiac remodeling, J. Exp. Med. 211 (8) (2014) 1673–1687.

[22] C.N. Serhan, N. Chiang, T.E. Van Dyke, Resolving inflammation: dual anti-in-flammatory and pro-resolution lipid mediators, Nat. Rev. Immunol. 8 (5) (2008)349–361.

[23] D. Kromhout, E.J. Giltay, J.M. Geleijnse, Alpha omega trial, n-3 fatty acids andcardiovascular events after myocardial infarction, N. Engl. J. Med. 363 (21) (2010)2015–2026.

[24] B. Rauch, R. Schiele, S. Schneider, F. Diller, N. Victor, H. Gohlke, M. Gottwik,G. Steinbeck, U. Del Castillo, R. Sack, H. Worth, H. Katus, W. Spitzer, G. Sabin,J. Senges, O.S. Group, OMEGA, a randomized, placebo-controlled trial to test theeffect of highly purified omega-3 fatty acids on top of modern guideline-adjustedtherapy after myocardial infarction, Circulation 122 (21) (2010) 2152–2159.

X. Fang et al. Prostaglandins and Other Lipid Mediators 139 (2018) 1–9

9