**1. Introduction**

Acute coronary syndrome (ACS) comprises a set of ischemic conditions including unstable angina (UA), myocardial infarction (MI) (with or without ST-segment elevation), and sudden cardiac death. It is the most common cause of morbidity and mortality worldwide, and accounts for roughly seven million deaths and 129 million loss of disability-adjusted life years (DALYs) annually [1]. The main cause of ischemia is the reduction of blood flow into coronary microcirculation as a result of atherosclerotic plaque rupture and thrombus formation [2]. Complete occlusion of coronary arteries usually presents with ST-segment elevation myocardial infarction (STEMI), which is accompanied by tissue injury and presents with elevated troponin levels. Partially occluded coronary arteries may result in non-STEMI or UA, depending on whether or not myocardial injury occurs [3].

Coronary angiography has shown that the atherosclerosis extent index (including the number of diseased vessels, stenosis and occlusions) is generally lower in ACS patients than in patients with stable angina, suggesting that plaque vulnerability rather than the extent of atherosclerosis may be the determinant of ACS [4]. The mechanisms leading to the progression of an asymptomatic plaque to a vulnerable one are not fully understood. A thin fibrous cap and a large lipid core (≥40% plaque volume), inflammatory cells, and high neovascularity are suggested as factors causing plaque vulnerability [5].

The oxidation of lipoproteins, namely oxidized low-density lipoproteins (Ox-LDLs), has been considered as a key factor in this transition through various mechanisms. Following the infiltration of LDL into the injured endothelium, LDL becomes oxidized to form Ox-LDL. This modified LDL elevates the expression of cell adhesion molecules such as intercellular cell adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), resulting in leukocyte (mainly monocytes and T-lymphocytes) recruitment and migration into the intima. In the intima, monocytes differentiate into macrophages. These lipid laden macrophages, which are called foam cells, along with the migrated T-lymphocytes release a variety of cytokines that promote inflammation and the generation of reactive oxygen species (ROS) [6]. Ox-LDL increases the infiltration of macrophages into the plaque (foam cell formation), up-regulates the expression of matrix metalloproteinase (MMP), and triggers proinflammatory reactions leading to plaque rupture [7].

Several clinical studies have confirmed that Ox-LDL concentrations are significantly higher in MI patients when compared with stable angina or age-matched controls [8–10]. Lipid peroxidation can occur within the LDL membrane through non-enzymatic and/or enzymatic mechanisms, producing diverse secondary products such as 4-hydroxynonenal (4-HNE), malondialdehyde (MDA), oxidized phospholipids (OxPLs), and oxylipins. These oxidized lipids are bioactive and can be bound to proteins, peptides, phospholipids, and nucleic acids, generating structural neo-epitopes called oxidation-specific epitopes (OSEs). Consequently, chronic elevations of OSEs may induce inflammation through the secretion of chemokines and proinflammatory cytokines, leading to plaque instability [11]. Clinical studies have also confirmed higher levels of these bioactive molecules in ACS patients when compared with patients in control groups [12–15].

Previous studies have shown that bioactive lipids can predict ACS occurrence in various populations. For instance, higher levels of OxPLs on Ox-LDL have been found to predict the progression of first or second major coronary events [16,17]. In addition, a recent cross-sectional study showed that levels of oxylipins, namely dihydroxy-eicosatrienoic acid (DiHETrE) and 16- hydroxy-eicosatetraenoic acid (HETE), were significantly associated with cardiovascular and cerebrovascular events, respectively. In this study, levels of 8,9-DiHETrE were significantly higher in patients with ACS (*n* = 24) compared to those without ACS (*n* = 74). Univariate and multivariate logistic regression also revealed that 8,9-DiHETrE concentrations were significantly associated with the presence of ACS. Moreover, they found that with every 1 nmol/L increase in the 8,9-DiHETrE concentrations, the odds of ACS increased by 454-fold. In this particular study, 8,9-DiHETrE elevated the odds of ACS by 92-fold [18].

Bioactive lipids have been measured conventionally by the use of chemical reagents, immunoassays, or chromatography [19]; however, these methods have limitations such as the lack of sensitivity and specificity. The main drawback of using conventional methods is that only one analyte can be assessed with one set of analysis. Considering the heterogeneity of pools of oxidized lipids, rapid multi-analyte quantification methods are needed. With the advent of robust mass spectrometric techniques, various groups of compounds can be assessed at the same time in a targeted and non-targeted fashion. By using soft ionization mass spectrometry (MS) such as electrospray ionization (ESI), the identification and quantification of non-volatile and thermolabile samples such as OxPL and oxylipins are feasible. Lipidomics is a powerful tool providing another layer of the detailed molecular levels of lipid assessments that may help to explore novel biomarkers and new treatment options in ACS [20].

In this article, we will briefly review the mechanisms in which bioactive lipids are generated. Then, we will focus on the analytical methods used by previous studies to measure these compounds. Finally, we will review the clinical studies that have assessed the roles of bioactive lipids in ACS patients.

#### **2. Bioactive Lipid Generation**

About 700 phospholipid (PL) molecules have been identified on the surface of LDL particles [6]. Phosphatidylcholine (PC) and sphingomyelin (SM) are the main PLs in LDL particles [21]. Most PLs

contain polyunsaturated fatty acids (PUFAs), with 14–24 carbons in their sn-2 position, which make them susceptible to oxidation. They can undergo non-enzymatic oxidation mainly by ROS, making heterogeneous pools of oxidized lipids. Hydroperoxides (LOOH) are the first products of PUFA oxidation by ROS. During degradation of LOOH, a large variety of secondary products are produced such as 4-hydroxynonenal (4-HNE), malondialdehyde (MDA), non-fragmented (full length), and fragmented (shorten chain) OxPLs [22] (Figure 1).

**Figure 1.** Non-enzymatic oxidation of membrane phospholipids. Free radicals may attack membrane phospholipids such as PAPC, leading to the production of bioactive lipid molecules. Abbreviations: PAPC-OOH, PAPC hydroproxide; OxPC, oxidized phosphatidylcholine; PEIPC, 1-palmitoyl-2- (5,6-epoxyisoprostane E2)-sn-glycero-3-phosphocholine.

4-HNE is a α,β-hydroxyalkenal which is formed through the peroxidation of arachidonic acid (AA) (20-carbon compounds) and linoleic acid (LA) (18-carbon compounds). Its reaction with the histidine, cysteine, or lysine residues of proteins makes Schiff bases or Michael adducts. MDA is a three-carbon aldehyde that is similarly produced through the non-enzymatic oxidation of PUFA. It can also be produced as a side product of thromboxane A2 (TXA2) synthesis. AA and docosahexaenoic acid (DHA) are the main precursors of MDA [23]. Levels of 4-HNE and MDA increase during oxidative stress and have been widely accepted as markers of oxidative stress.

OxPLs can be divided into two groups of non-fragmented (with the same number of carbon with precursor) and fragmented (with shorter chain) OxPLs. Non-fragmented OxPLs are formed following the initial phase of lipid oxidation. Then, they may undergo intramolecular cyclization, rearrangement, and further oxidation and make OxPLs with terminal furans, isoprostanes, and long-chain products with functional groups such as hydroperoxides, hydroxides, keto- and epoxy-groups [24]. Fragmented OxPLs have hydroxide and carbonyl groups in their structures, which are highly bioactive and can rapidly interact with biomolecules causing tissue injury [25] (Figure 2).

**Figure 2.** Fragmented and non-fragmented OxPC productions from PAPC. OxPLs can be classified as fragmented and non-fragmented species. Non- fragmented species are produced from the addition of peroxyl radicals where rearrangement/cyclization may happen. Fragmented species are comprised of aldehyde and carboxylic acid containing lipids. Abbreviations: Oxo-ETE-PC, oxoeicosatetraenoic acid phosphocholine; PEIPC, 1-palmitoyl-2-(5,6-epoxyisoprostane E2)-sn-glycero-3-phosphocholine; Aldo-OxPC, aldehyde containing oxidized phosphatidylcholine; Keto OxPC; carboxylic acid containing oxidized phosphatidylcholine; POVPC, 1-palmitoyl-2-(5 -oxo-valeroyl)-sn-glycero-3-phosphocholine; PGPC, 1-palmitoyl-2-glutaryl-sn-glycero-3-phosphocholine.

All PUFAs including omega-3 PUFAs are oxidized by the three main enzymes of cyclooxygenase (COX), lipoxygenase (LOX), and cytochrome P450 (CYP). The types of oxylipins produced from the PUFAs depend on the type/amounts of dietary PUFA, and the availability and affinity of the enzymes (COX, LOX, or CYP) for a specific substrate PUFA. The most well-known oxylipins are derived from AA and LA [1]. Half of the known oxylipins are derived from AA. However, other oxylipins can also be produced from PUFAs besides AA including both the omega-3 and omega-6 PUFA. It is important to mention that phospholipase-A2 (PL-A2), which has a key role in oxylipin production, has a preference for AA and eicosapentaenoic acid (EPA) [26]. These fatty acids may undergo enzymatic oxidation through cyclooxygenase (COX), lipoxygenase (LOX), and cytochrome P450 (CYP) pathways. Oxylipins are not stored in the cells and exert their biological roles through paracrine or autocrine mechanisms [27] before they are chemically inactivated or re-esterified into a glycerolipid pool [28] (Figure 3).

**Figure 3.** Enzymatic oxidation of membrane phospholipids. Fatty acids are released from the membrane PL by the phospholipase A2 enzyme and may undergo oxidation through three oxidation pathways including COX, LOX, and CYT P450. Abbreviations: COX, cyclooxygenase; LOX, lipoxygenase; CTY P450, cytochrome P450; HETE; hydroxyeicosatetraenoic acids; EET, epoxyeicosatrienoic acids.

Prostanoids (prostaglandines (PG) and thromboxanes (TX)) and some forms of hydroxy-metabolites such as 11-HETE are generated through the COX pathway from AA. LOX enzymes catalyze the generation of hydroxy fatty acids such as leukotrienes, lipoxins, resolvins, protectins, maresins, hepoxilins, and eoxins [3,29]. Mid chain (5-, 8-, 9-, 11-, 12-, and 15-) HETEs are also formed from AA through the LOX pathway [18,30]. CYP 450 enzymes have epoxygenase or ω-hydroxylase activity [29]. ω-terminal (16-, 17-, 18-, 19-, and 20-) HETEs are produced from AA and by ω-hydroxylase enzymes (CYP4A and CYP4F) and epoxyeicosatrienoic acid (EETs) are generated by CYPs with epoxygenase activity [28].

#### **3. Measurement of Bioactive Lipids**

#### *3.1. 4-HNE and MDA*

Free aldehydes can be identified and quantified by several analytical methods. Thiobarbituric acid reactive substance (TBARS)/spectrophotometry has been widely employed to measure MDA levels. Under acidic conditions and high temperatures, the aldehyde group of MDA reacts with the nucleophilic

center of TBA and makes a red-colored derivative, which can be detected by spectrophotometric and spectrofluorometric approaches. The aldehyde group of HNE can also make derivatives with 2,4-dinitrophenylhydrazine (DNPH) that are detectable by spectrophotometry [12,14,31,32].

Kami ´nski et al. (2008) measured the HNE and MDA in the plasma of 15 STEMI patients and 10 patients with stable IHD as the control group by using derivatization/high performance lipid chromatography (HPLC)-fluorescence detection [33]. Solid phase extraction was applied to extract HNE and MDA [34]. MDA was also detected using the TBARS derivatization, and then separated and quantified by HPLC-spectrofluorometric assay [35].

Gas chromatography (GC) is the other main analytical method to measure MDA and HNE. MS is more specific and sensitive compared with other analytical methods as it can identify these aldehydes based on the mass to charge ratio and fragmentation pattern [36]. GC can also be coupled to MS. Tsikas et al. (2017) developed a method to measure the plasma concentrations of both MDA and HNE simultaneously by using GC/MS. They used pentafluorobenzyl hydroxylamine as a derivatization reagent, [1,3-2H2]-MDA (d2-MDA), and [9,9,9-2H3]-HNE (d3-HNE) as the internal standards. The ionization technique used here was hard ionization with high energy such as electron impact [37] and is different from soft ionization, which will be discussed later.

Syslová et al. (2009) developed a method using reverse phase HPLC/ESI-MS to assess the MDA and HNE in plasma, urine, and exhaled breath condensate [38]. HNE-d3 and Me-MDA was used as the internal standards, butylated hydroxytoluene (BHT) as the antioxidant, and acetonitrile was added to the plasma. Then, the plasma was sonicated and centrifuged to remove the precipitated proteins. The supernatant was dried under nitrogen gas, and re-suspended in acetonitrile to be injected into HPLC. HPLC with a Hypercarb Thermo 100 mm × 2.1 mm × 5 mm column and Hypercarb-precolumn was used. Water and ammonium hydroxide were used as solvent A, and methanol:acetonitrile with ammonium hydroxide was used as the co-solvent. Derivatization with 4-2-trimethylammonio ethoxy benzenaminiumhalide (4-APC) or cyclohexanedione (CHD) can be also done prior to extraction to increase the ionization of these aldehydes [39,40].

#### *3.2. OxPL*

Using monoclonal antibodies is one of the well-established methods to assess the OxPL levels on apoB100-containing lipoproteins, namely LDL, very low-density (VLDL), and lipoprotein (a) (LP (a)). To perform this assay, the murine monoclonal antibody MB47 must be added initially to capture all apoB-100 lipoproteins from the plasma. Then, by adding the E06 antibody, it can bind to apoB-100. This method has been applied extensively to measure the OxPL levels in CVD [41]. The limitation of this method is that only the OxPL species that are present on the apoB100 lipoproteins can be assessed, and not the total amount of OxPL in the plasma. In addition, this method cannot identify specific OxPL species among all types of OxPL (fragmented, non-fragmented OxPL) that are produced during lipid oxidation [19]. To overcome these limitations, LC/MS has been introduced as the best option for a detailed analysis of OxPLs.

Hassanaly et al. (2017) measured the levels of oxidized phosphatidylinositol (OxPI) in Ox-LDL and human atherosclerotic plaque by using reversed-phase HPLC/ESI-MS. Using this approach, they were able to identify and quantify 23 OxPI species in human Ox-LDL and atherosclerotic plaque. They found that levels of OxPI species increased significantly in Ox-LDL at 48 h when compared with the baseline. Moreover, non-fragmented hydroperoxides were the dominant species in Ox-LDL at 48 h, comprising 52.07% of the total OxPI species. Fragmented aldehyde and carboxylic acid containing OxPI comprised 17.32% and 0.89% of total the OxPI at the same time point. Likewise, in human atherosclerotic plaques, which were retrieved from patients who underwent saphenous vein graft (SVG) interventions, non-fragmented hydroperoxides were the most abundant OxPI compounds. Fragmented aldehyde and carboxylic acid containing OxPI comprised 18.6% and 1.5 % of the total OxPI compounds, respectively [42].

OxPC species have been identified in patients that have undergone percutaneous coronary and peripheral procedures by using normal phase HPLC/ESI-MS. In this study, the five most abundant OxPCs were in embolized material captured by distal protection filter devices during uncomplicated saphenous vein graft, carotid, renal, and superficial femoral artery interventions. 1-palmitoyl-2-(9-oxo-nonanoyl) PC (PONPC) was the most abundant fragmented OxPC, which comprised 50% of the total quantified fragmented OxPC compounds. POVPC, 1-palmitoyl-2-glutaroyl-sn-glycero-3-phosphocholine (PGPC), and 1-palmitoyl-2-(5-keto-6-octene-dioyl) PC (KOdiAPC) were the other fragmented OxPC species measured in this study [43].

Recently, we were able to identify and quantify 56 OxPC species including both fragmented and non-fragmented OxPCs in rat kidneys following ischemia/reperfusion (I/R) injury. 1-stearoyl-2-linoleoyl-phosphatidylcholine (SLPC-OH) and 1-palmitoyl-2-azelaoyl-sn-glycero-3 phosphocholine (PAzPC) were the most abundant non-fragmented and fragmented OxPC after I/R, respectively. The total levels of OxPC species (including fragmented and non-fragmented OxPC compounds) increased significantly after both 6 h and 24 h reperfusion when compared with the sham group. Concentrations of fragmented OxPCs were elevated significantly by increasing the time of reperfusion as their levels were significantly higher following 24 h reperfusion when compared to 6 h I/R and sham groups. However, no significant differences were observed between the sham and 6 h I/R groups. Changes in the levels of non-fragmented OxPCs were different to the fragmented compounds. Although the total levels of non-fragmented OxPC elevated significantly in the 6 h I/R group, no differences were observed in the 24 h I/R group. These data pointed to the importance of identifying the specific compounds, and not just the total concentrations of the oxidized species [44].

The first step in preparing samples for lipidomic analysis is extracting the lipids from the cell/tissue/plasma. Currently, conventional liquid–liquid extraction has been widely used for the extraction of OxPL [45]. Folch extraction, which uses chloroform/methanol, is one of the most common extraction approaches to extract OxPL. Adding antioxidants such as BHT is recommended to minimize further oxidation [46]. Recently, it has been suggested that enrichment strategies such as using gold nanoparticles (GNP) and anti-Ox-LDL antibodies on plasma samples [47] or lipid extracts [48] may increase the efficacy of OxPC identification. Hinterwirth et al. [47] used GNPS with four different Ox-LDL antibodies, namely the E06, anti-Cu Ox-LDL antibody, anti-MDA-LDL antibody, and anti-carboxymethyllysine-LDL antibody, to increase the detection of OxPC in plasma. Stübiger et al. [48] also reported that using 2-aminobenzoic acid (2-AA) as the reagent with GNP elevated the carbonyl-containing OxPC identification at subnanomolar concentrations, with up to 90% recoveries [49].

To separate the OxPLs species, reverse phase HPLC with C8 or C18 columns with either isocratic or gradient elutions has been widely used, although they can also be separated by normal phase HPLC [43]. By using HPLC, OxPL are separated based on polarity and molecular weight before interfacing with the MS, which increases the sensitivity of the assessment [50]. Reverse phase mobile phases are usually a mixture of water, methanol, or acetonitrile. Hexane or isopropanol can also be applied as co-solvents. Ammonium acetate, ammonium formate, or acetic or formic acid may also be added to the solvent to facilitate ionization in MS. There are no detruded internal standards for OxPLs analysis. Non-oxidized PLs and lyso PL (LPL) such as phosphatidylinositol (PI) (31:1) for OxPI analysis and PC (9:0)/LPC(17:0) for OxPC analysis have been used as internal standards [42,44]. These PLs have the same structures and fragmentation patterns and are not produced in the body. Therefore, they can be used as an internal standard to assess the extraction efficacy and instrument response [51].

ESI/MS and matrix-assisted laser desorption/ionization (MALDI) are two forms of soft ionization techniques. The soft ionization technique allows for the analysis of non-volatile compounds such as OxPL. ESI can readily interface with HPLC. This is very important when analyzing OxPLs as the levels of these compounds are considerably lower when compared with non-oxidized compounds. Therefore, separation techniques prevent ion suppression, which may occur with high abundant

molecular ions. On the other hand, sample preparation is simpler with ESI when compared with MALDI as MALDI needs the co-crystallization of a matrix with the sample, which consequently may affect the quantification of the analytes. MALDI can examine solid state samples and is useful for MS imaging of tissue, while ESI needs tissue extraction as it requires a liquid sample [52]. Some studies have used MALDI to quantify chlorinated PL [53,54]. However, no study has measured the levels of oxidized lipids in tissue using MALDI. In a review by Ana Reis (2017), it was emphasized that MS imaging to assess the distribution of OxPL in tissue is challenging due to the low concentrations of OxPL/PL and the lack of fluorescent probes designed to bind to free OxPL in tissue samples [45].
