Variability in the Clinical Effects of the Omega-3 Polyunsaturated Fatty Acids DHA and EPA in Cardiovascular Disease—Possible Causes and Future Considerations

Abstract

Cardiovascular disease (CVD) that includes myocardial infarction and stroke, is the leading cause of mortality worldwide. Atherosclerosis, the primary underlying cause of CVD, can be controlled by pharmacological and dietary interventions, including n-3 polyunsaturated fatty acid (PUFA) supplementation. n-3 PUFA supplementation, primarily consisting of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), has shown promise in reducing atherosclerosis by modulating risk factors, including triglyceride levels and vascular inflammation. n-3 PUFAs act by replacing pro-inflammatory fatty acid types in cell membranes and plasma lipids, by regulating transcription factor activity, and by inducing epigenetic changes. EPA and DHA regulate cellular function through shared and differential molecular mechanisms. Large clinical studies on n-3 PUFAs have reported conflicting findings, causing confusion among the public and health professionals. In this review, we discuss important factors leading to these inconsistencies, in the context of atherosclerosis, including clinical study design and the differential effects of EPA and DHA on cell function. We propose steps to improve clinical and basic experimental study design in order to improve supplement composition optimization. Finally, we propose that understanding the factors underlying the poor response to n-3 PUFAs, and the development of molecular biomarkers for predicting response may help towards a more personalized treatment.

1. Introduction

1.1. Cardiovascular Disease

Cardiovascular disease (CVD), including myocardial infarction and ischemic stroke, is the leading cause of death worldwide. Atherosclerosis, which entails the formation of an atheromatic plague due to build-up of lipids and fibrous tissue in the large arteries, is the primary underlying cause of CVD [1]. Stenosis caused by the atheromatic plague, as well as thrombosis due to plague rupture, can restrict blood flow in major vessels leading to myocardial infarction or ischemic stroke [2].

Myocardial infarction, caused by myocardial ischemia, involves myocardial necrosis, as well as other cell death-activating pathways due to oxygen deprivation [3]. The main pathological cause of myocardial infarctions is associated with coronary artery disease, which involves coronary artery blockage due to accumulation of atherosclerotic plagues [4]. Myocardial damage can eventually lead to heart failure, which involves anatomical and functional myocardial abnormalities that limit ventricular filling or blood ejection, and results in failure to satisfy the underlying needs of circulation [5]. Ischemic stroke involves a reduction in the blood flow to part or all of the brain, leading to tissue damage and some degree of neurological damage. The major cause of ischemic stroke is the rupture of atherosclerotic plagues and the formation of thrombosis in the carotid artery [6].

1.2. Atherosclerosis

Atherosclerosis is primarily a lipid-driven process caused by accumulation of low-density lipoprotein (LDL) and remnant lipoprotein particles in focal areas of arteries, and an associated inflammatory process particularly at regions of disturbed non-laminar flow at artery branch-points. The defined most frequent risk factors are: high LDL cholesterol (LDL-C), hypertension, diabetes mellitus, smoking, age, and family history. Sedentary lifestyle, obesity, diets high in saturated and trans fats, and specific genetic abnormalities may also increase the risk [7,8,9].

Atherosclerosis primarily develops due to trapping of lipids in the intima by the extracellular matrix, causing a modification that drives chronic inflammation at vulnerable sites in the arteries. This process, which is involved in all stages of atherogenic progression, starts with nascent fatty streaks in the artery intima, which develop into fibrous plaques and then into rupture-prone atherosclerotic lesions [10,11]. At the molecular level, LDL in the intima becomes oxidized and drives pro-inflammatory processes that attract monocytes to the artery wall, where they differentiate into macrophages. Macrophages take-up lipoproteins using LDL scavenging receptors, forming foam cells that create the atherosclerotic lesion. Inflammatory mediators also recruit vascular smooth muscle cells, which produce extracellular matrix proteins creating a fibrous cap that overlies the lesion, forming an atherosclerotic plague [12]. Reduced synthesis of extracellular matrix proteins by smooth muscle cells, and increased production of matrix metalloproteases by foam cells may lead to fibrous cap thinning and plague rupture leading to thrombosis [13,14,15]. In advanced plagues, there is increased smooth muscle cell and macrophage cell death, through apoptosis or necroptosis [16,17]. Failure of macrophages to clear apoptotic and necrotic cells by efferocytosis, leads to the formation of the necrotic core that precipitates inflammation and fibrous cap thinning [12,18].

Treatment for coronary atherosclerosis involves measures to encourage regression and stop the growth and rupture of atherosclerotic plaques. Treating risk factors such as elevated LDL-C, high blood pressure, and diabetes, through diet, exercise, smoking cessation, and pharmacological management are the key strategies for controlling atherosclerosis [19,20,21]. Statins are the mainstay for lowering LDL-C and preventing cardiovascular events and mortality. Angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, diuretics, beta-blockers, calcium channel blockers, and vasodilators are also used for blood pressure management [22,23,24].

1.3. Dietary Management of Atherosclerosis

Atherosclerosis can be controlled by a healthy diet that is rich in fibre, monounsaturated fats, oily fish, fruits, and vegetables, and low in saturated and trans fats [25]. n-3 polyunsaturated fatty acids (PUFAs) have received considerable attention for their potential in modulating key risk factors of CVD, including triglyceride levels, lipoprotein oxidation, vascular inflammation and thrombogenesis. Nonetheless, clinical studies investigating the effects of n-3-PUFAs on cardiovascular health show conflicting findings [26]. A better understanding of the factors underlying the variability in the effects of n-3-PUFAs in clinical studies will allow us to tailor supplementation regimes in order to achieve greater benefit for the patients. In this review, we will discuss possible causes of the reported variability, by focusing on atherosclerosis. Furthermore, we provide suggestions for improving pre-clinical and clinical studies in order to better understand and optimize n-3 PUFA supplementation for preventing CVD.

2. Current Status of Knowledge

2.1. Biochemistry and Structural Morphology of n-3 PUFAs

n-3 (omega-3) PUFAs are a family of long chain cis polyunsaturated fatty acids [27]. The term n-3 reflects the fact that the first double bond is situated three carbon atoms from the methyl terminal group [27]. Alpha linolenic acid (ALA; C18:3) is precursor to the longer chain (LC) n-3 PUFAs, eicosapentaenoic acid (EPA; C20:5) and docosahexaenoic acid (DHA; C22:6) [28]. ALA is ‘essential’ and must be included in our diet since it cannot be synthesised de novo in the body. Oily fish are the main source of LC n-3 PUFAs [29]. n-6 (omega-6) FAs are the second known important family of PUFAs. Linoleic acid (LA; C18:2) is a precursor to the LC n-6 PUFA arachidonic acid (C20:4). The main sources of n-6 PUFAs include the common vegetable oils used in cooking, such as sunflower and soybean oil, as well as foods derived from livestock animals and poultry [30].

It has been widely accepted that the present Western diet is ‘deficient’ in n-3 PUFAs with a ratio of n-6 to n-3 of 15-20:1, which is far higher than the optimal 4:1 ratio and the ideal 1:1 ratio [31]. However, reducing n-6 PUFA intake is not a prerequisite to achieve the optimum ratio. According to Zhao et al., n-6 PUFA intake can be increased without inducing any adverse effect provided that adequate n-3 PUFAs are consumed. Moreover, sufficient consumption of n-6 PUFAs is important for reducing LDL-C concentrations and therefore, a sufficient intake of both n-3 and n-6 PUFAs is crucial for CVD risk reduction [32].

2.2. Clinical Evidence on the Effects of n-3 PUFAs on Different CVDs

Epidemiological studies have shown a positive correlation between oily fish intake and beneficial cardiovascular effects [33]. The evidence of the benefits of oily fish is stronger in secondary than in primary prevention settings [34]. For instance, GISSI Prevenzione investigators (1999) showed that dietary supplementation with LC n-3 PUFAs (1 g/day) in patients after myocardial infarction reduced total mortality by 21% and sudden death by 45%. In addition, a recent pooled analysis of four cohort studies showed an association between oily fish intake and lower risk of mortality among patients with prior CVD [35]. These benefits have been attributed to the LC n-3 PUFAs, EPA and DHA, primarily found in oily fish [36]. ALA is generally far less effective at inducing biological effects, partly because of its inefficient conversion (<5%) to EPA and DHA in humans [28]. ALA is inefficiently converted into the LC n-3 PUFAs partly because of the large and increasing amounts of n-6 fatty acids present in our diet, which compete for the same enzymes, shown in the metabolic pathway [37,38] (Figure 1).

The beneficial effect of the LC n-3 PUFAs on reducing the risk of cardiac mortality appears to be due to their incorporation into cardiomyocyte phospholipids at the expense of arachidonic acid [39,40]. The presence of n-3 PUFAs, particularly EPA, in the cell membranes can lead to the generation of different eicosanoids which can be more cardioprotective than those resulting from the arachidonic acid cascade [41,42,43]. Several potential mechanisms for the cardioprotective effect of omega-3 fatty acids have been proposed with the effect on blood lipids, such as triglycerides (TG), LDL, and HDL, associated with risk of atherosclerosis, being the most scientifically proven [44,45]. Particularly, the most consistent effect of n-3 PUFAs is the reduction in serum TG [46].

There is a substantial confusion among the public and health professionals regarding the overall effects of n-3 PUFAs on blood lipids [47]. It has been suggested that elevated TG levels are independent risk factors for the progression of CVD [48]. In addition, smaller diameter LDL and HDL particles have been associated with increased risk [49,50]. The ‘Omega-3 index’ [51], which describes the content of EPA and DHA in red blood cell (RBC) membranes, expressed as a proportion of total FAs, has been considered a risk factor for death from coronary heart disease (CHD) [52].

2.2.1. Effect on Triglycerides

According to Milte et al., circulating TG levels constitute an independent risk factor for CVD and are correlated with the severity and development of atherosclerosis [53]. However, Torrejon et al., argued that unlike the well-established role of LDL-C in the development of CVD, the role of circulating TG concentrations in CVD development remains controversial [54]. Nevertheless, according to the British Nutrition Foundation (BNF) (2005) [55], an indirect effect of TGs on CVD risk seems to be the case. Research has shown that a reduction in TG levels leads to a lower abundance of small, dense LDL-C and therefore a reduced CVD risk (BNF, 2005) [55].

The most consistent effect of n-3 PUFAs is the decrease in fasting and postprandial serum TGs [48,54]. However, the exact dose as well as exact duration of intervention, which can give the optimum TG-lowering effect is still unclear [53]. So far, the majority of the studies have demonstrated a significant reduction in TGs (25–30%) following treatment with doses ≥ 3 g/day of LC n-3 PUFAs (EPA and DHA) mainly in the form of fish oil [56]. In contrast, there is evidence showing a much lower effect following treatment with lower doses [57]. The duration of studies ranged between 6 and 104 weeks, and the optimal duration of intervention remains unclear since no difference was found in TG levels associated with longer trial duration (>16 weeks) compared to shorter trial duration (≤16 weeks) [33,58].

A double-blind randomized placebo controlled parallel study showed that DHA supplementation (3 g/day) for 45 days significantly decreased fasting TG levels of hypertriglyceridaemic men by 25–30%. Conversely, a double-blind placebo-controlled study [57] using a LC n-3 PUFA dose (1.8 g EPA and 0.3 g DHA) per day, which approximated the daily dietary intake upper limit (2 g of LC n-3 PUFAs) of the current UK guideline range (Scientific Advisory Committee on Nutrition, 2004) [59], had no effect on TG levels of normotriglyceridaemic individuals. Yusof et al. suggested that apart from the lower dose, the fact that the oil used was relatively poor in DHA might have contributed to the lack of effect. It is now believed that DHA is more potent in lowering TGs than EPA [60]. Furthermore, it has been shown that the DHA TG-lowering effect is greater in hypertriglyceridaemic individuals compared with normotriglyceridaemic individuals [61,62]. A comprehensive meta-analysis (16511 participants in 47 studies included) assessing the role of n-3 PUFAs in treating hyperlipidaemia, demonstrated an average reduction of 14% in TG levels of hypertriglyceridaemic individuals following 6-month treatment with an average daily intake of 3.25 g of EPA and/or DHA [33]. However, this dose exceeds the safety limit (3 g/day) approved by the US food and drug administration (FDA) [63].

2.2.2. Effect on Low-Density and High-Density Lipoprotein Particle Size

Smaller diameter and denser LDL particles, such as LDL-3 subfraction, have been shown to be more susceptible to oxidation [50]. In addition, they have an increased ability to penetrate the intima than larger, less dense LDL particles, such as LDL-1 and LDL-2 subfractions [50] and have, therefore, been associated with an increased coronary heart disease risk [50]. A very recent systematic review concluded that small dense LDL are associated with an increased CVD risk [64]. According to Torrejon et al., particle size is significantly increased with DHA supplementation. DHA supplementation reduces plasma TGs, contributing to a reduction in the number of small, dense LDL particles and subsequently to decreased CVD risk [49].

Larger, more cholesterol-rich HDL particles (e.g., HDL-2 subfraction) are thought to facilitate a more efficient reverse cholesterol transport compared with the smaller, less buoyant subfraction HDL-3, making them more cardioprotective [50,65,66]. The HDL-2 subfraction has been shown to be increased in response to DHA, whilst it is not affected by EPA [54,65]. This was confirmed by a recent systematic review [67].

2.2.3. Effect on ‘n-3 Index’

Harris and von Schacky (2004) were the first to propose that the content of EPA and DHA in RBC membranes, expressed as a proportion of total FAs, termed the ‘omega-3 index’, can be considered as a risk factor for coronary heart disease and sudden cardiac death mortality [51]. Specifically, an ‘omega-3 Index’ level of ≥8% is a reasonable preliminary target value for reducing coronary heart disease risk. On the other hand, an ‘omega-3 Index’ < 4% has been implicated with a 10-fold greater risk of sudden cardiac death compared with an ‘omega-3 Index’ of ≥8% (Figure 2) [68,69].

An increased proportion of the LC n-3 PUFAs (EPA and DHA) in erythrocyte membranes following consumption of oily fish has been demonstrated [70]. A randomised, double-blind, placebo-controlled parallel trial investigated the association between increases in erythrocyte DHA content and changes in blood lipids in 67 healthy individuals following DHA intake for 3 months [53]. Participants were randomised to receive 0.52 g, 1.04 g, 1.56 g DHA or 1 g Sunola oil (control). At the end of the 3-month intervention the 0.52 g, 1.04 g, 1.56 g DHA doses increased the proportion of DHA in erythrocytes to a dose-dependent manner (7.1%, 7.9% and 9% of total FAs, respectively) [53].

3. Factors Implicated in the Efficacy of PUFAs

3.1. Variability in the Clinical Effects of n-3 PUFAs

Following supplement uptake, n-3 PUFAs are incorporated into cell membranes and plasma lipids in a dose- and time-dependent manner, replacing other fatty acid types, including n-6 PUFAs [71]. n-3 PUFA incorporation lead to changes in cell signaling by affecting the cell membrane structure and fluidity and changing the function of cell surface receptors and ion channels. The effects of n-3 PUFAs on atherosclerosis are conferred through molecular mechanisms that involve changes in the cell membrane composition, eicosanoid synthesis, transcription factor activity and gene expression [72]. EPA and DHA are known to have differential metabolism, and divergent effects on the molecular mechanisms of atherosclerosis. Antagonism between the two types of n-3 PUFAs, may explain the conflicting findings of large clinical studies on the effects of n-3 PUFAs on CVD risk factors [73,74,75,76]. The effects of n-3 PUFAs on atherosclerosis risk factors depend on different considerations including their dose and length of supplementation, EPA/DHA composition and formulation [77]. Understanding how these factors affect their incorporation into the lipid pool, their metabolism and their downstream molecular effects will allow more effective design of n-3 PUFA supplementation strategies for CVD prevention.

3.2. Incorporation of EPA and DHA into the Lipid Pool

EPA and DHA show differential metabolism-, tissue-, and compartment-specific accumulation following supplementation, which may affect their function. Supplementation with EPA and DHA leads to a dose-dependent increase in plasma, which reaches equilibrium approx. one month post-supplementation [78]. Studies suggest a very low conversion of EPA to DHA following supplementation. However, DHA supplementation increases plasma EPA concentrations, possibly due to retro-conversion or slow EPA turnover [78,79]. Following fish oil supplementation, EPA preferentially accumulates in cholesteryl esters within very low-density lipoproteins (VLDL), possibly due to increased selectivity of the enzyme lecithin-cholesterol acyltransferase to EPA [80,81,82]. On the other hand, DHA is a preferential substrate for diacylglycerol acyltransferase leading to increased incorporation into TG [81,82,83].

The source and formulation of n-3-PUFA supplements also affects their bioavailability and consequently their lipid incorporation. Studies comparing the incorporation of EPA and DHA into plasma lipids after short-term supplementation with nutritional or pharmacological doses (>3 g/day) of n-3-PUFAs, in the forms of free fatty acids, ethyl esters or re-esterified TG, have reported conflicting findings [84,85,86,87]. However, a clinical study investigating the effect of prolonged supplementation (6 months) of dyslipidemic patients on statins with moderate doses of n-3-PUFAs (1.01 g EPA and 0.67 g DHA), reported that the re-esterified triglyceride formulation showed higher incorporation into red blood cell membranes and a more effective reduction in fasting plasma TG, compared to the ethyl ester formulation [88,89]. The above study reiterates the importance of supplement formulation and length of treatment on its effectiveness.

3.3. Effects of EPA and DHA on Lipoprotein Metabolism

High levels of blood TG, transported by the TG-rich VLDLs and chylomicrons, increase the risk for atherosclerosis [90]. Dietary supplementation with fish oil or with pure EPA and DHA, attenuates plasma TG by inhibiting TG synthesis and VLDL production, and inducing apolipoprotein B degradation in hepatocytes [91,92,93,94]. Furthermore, n-3-PUFAs limit the supply of plasma non-esterified fatty acids for VLDL synthesis through inhibition of intracellular lipolysis in adipocytes [95,96]. Accelerated chylomicron clearance through induction of lipolysis has also been shown to be induced by both DHA and EPA [97]. A number of studies also demonstrate that n-3-PUFAs supplementation increases HDL levels; however, there are studies showing contradicting findings [65,98,99,100].

Studies using high doses of purified EPA and DHA, given separately or together, alone or in combination with statins, show a significant improvement of dyslipidemia [89,101,102]. Skulas-Ray et al., reported that supplementation with high doses (3.4 g/d) of EPA and DHA for eight weeks led to a significant reduction in plasma TG, whilst a lower dose (0.85 g/d) had no significant effect. Neither of the doses had any effect on endothelial function and inflammatory markers, possibly because the study subjects were healthy [102]. These findings suggest that high doses of purified n-3 PUFAs may be required to achieve optimum clinical efficacy.

However, a number of studies using EPA or DHA monotherapy report differential effects or efficacies of the two fatty acids on plasma lipids. The randomised cross-over study ComparED, compared the effect of DHA and EPA (2.7 g/day), formulated as re-esterified triacylglycerol, on blood lipids and inflammatory markers in subjects with abdominal obesity and subclinical inflammation. The study reported that DHA may be more effective in reducing TG, and increasing HDL- and LDL-cholesterol concentrations, compared to EPA. Apolipoprotein B levels were also reduced by DHA, but not by EPA [103]. A secondary analysis of the ComparED study results indicated that the increased efficacy of DHA in reducing TG does not depend on a greater effect magnitude but on the fact that a greater proportion of study subjects responded to DHA compared to EPA [104]. Previous studies also investigated the effect of supplementation with highly pure, free, or esterified, DHA or EPA (>3 g/day) for 4–7 weeks on the plasma lipid profile of healthy or hypertensive subjects. These studies report mixed findings, with some showing a superior effect of DHA compared to EPA supplementation and others no difference between the two n-3 PUFAs [97,105,106,107,108]. A meta-analysis of these studies concluded that overall DHA reduces TG and increases LDL and HDL cholesterol, to a greater extent than EPA [109]. High doses (3–4 g/day) of EPA or DHA were shown to be equally efficient in inducing liposome lipase activity, suggesting that the superior triglyceride-lowering effect of DHA is not due to a more efficient chylomicron/VLDL clearance [97,110].

Early studies have demonstrated that n-3 PUFAs reduce TG levels, at least partly, through inhibition of acyl CoA: diacylglycerol acyltransferase (DGAT), an enzyme catalysing the terminal step of TG synthesis [111]. n-3 PUFAs also act by inhibiting hepatic de novo synthesis of fatty acids and TG and inducing fatty acid oxidation and TG catabolism in adipose tissue and muscle [112]. These effects are mediated by modulating the activity of transcription factors, and more specifically the sterol regulatory element-binding protein (SREBP)-1 and the peroxisome proliferator-activated receptors (PPARs) [72]. SREBPs are cellular fatty acid sensors, which are activated by proteolytic release from the endoplasmic reticulum leading to their translocation to the nucleus, where they activate lipogenic gene transcription. SREB-1a and -1c isoforms are transcribed from the SREBP-1 gene under the regulation of different promoters [113]. PUFAs, including n-3 PUFAs, inhibit SREBP-1 activation by preventing its proteolytic release from the ER, and reducing its gene expression by competing with its transcriptional activator heterodimer liver X receptor/retinoid X receptor (LXR/RXR) or by increasing its mRNA degradation [114,115,116]. RXR also form heterodimers with the nuclear receptors PPARs, which act as lipid sensors by binding to a number of different intracellular fatty acids species. PPARα, predominantly expressed in the liver, and PPARβ/δ that show higher expression in the skeletal and cardiac muscle, drive mitochondrial and peroxisomal fatty acid oxidation gene transcription. PPARγ is mainly found in adipose tissue where it regulates adipocyte differentiation and activity, including lipoprotein lipase activity and fatty acid oxidation. n-3 PUFAs act as ligands of RXR and all three PPAR isoforms, activating fatty acid oxidation and lipoprotein lipase gene expression [117,118,119].

Evidence from in vivo and in vitro studies suggests that EPA and DHA may differentially regulate these molecular transcriptional pathways. Studies in rat hepatocytes have shown that EPA is a more effective inhibitor of DGAT activity, compared to DHA. Furthermore, although EPA and DHA activate PPARα to a similar extent, EPA activates mitochondrial fatty acid oxidation, whilst DHA drives only peroxisomal fatty acid oxidation [83,120,121]. This is a result of differential substrate preference of each pathway, with EPA being oxidised both in microsomes and mitochondria, whereas DHA is only oxidised in microsomes [83] Based on these experimental findings, EPA may have a greater influence on TG levels, as mitochondrial fatty acid oxidation has a more significant influence on fatty acid availability for triglyceride synthesis, compared to microsomal oxidation [122]. Unfortunately, to the best of our knowledge, there are no published studies comparing the effects of EPA and DHA on fatty acid oxidation in humans.

Studies in rat models have shown that EPA and DHA inhibit lipogenesis both separately and in combination [123]. EPA has also been shown to inhibit LXR/RXR-SREBP-1c interaction more potently than DHA, in a human embryonic kidney cell line [114]. However, a randomised control trial, comparing the effects of DHA and EPA (3 g/day) on plasma triglyceride metabolism in young healthy subjects, reported that although both n-3-PUFAs inhibited liposome lipase activity, EPA increased the lipogenic index and failed to reduce plasma triglyceride levels. In contrast, DHA did not modulate the lipogenic index and significantly reduced triglyceride levels [110].

4. Specific PUFAs as Anti-Inflammatory and Anti-Oxidative Mediators

4.1. Effects of EPA and DHA on Inflammation

Another way through which n-3-PUFAs exert their protective effect against atherosclerosis, is by reducing vascular inflammation. Randomised control trials show that a mix of DHA and EPA in different proportions and formulations reduces plasma inflammatory markers, including C-reactive protein (CRP), IL-6 and TNFα, and increases the anti-inflammatory mediator adiponectin [124,125]. A recent meta-analysis of studies directly comparing the effects of DHA and EPA on plasma inflammation concluded that the two n-3-PUFAs do not have significantly differential effects on CRP, IL-6, TNFα and adiponectin [126].

n-3-PUFAs exert their anti-inflammatory effect by modulating the balance between pro- and anti-inflammatory lipid mediators. Following their release from phospholipids by phospholipases, PUFAs are metabolised by cyclooxygenases (COX), lipoxygenases (LOX) and cytochrome P450 (CYP450) into oxylipins that can act as endogenous mediators. n-6-PUFAs, such as arachidonic acid, act as pre-cursors for pro-inflammatory and pro-thrombotic eicosanoids, such as the 2-series prostaglandins and thromboxanes, and the 4-series leukotrienes. EPA gives rise to 3-series prostaglandins and thromboxanes and 5-series leukotrienes, which are less inflammatory. Furthermore, EPA and DHA promote inflammation resolution by being metabolised into anti-inflammatory mediators, such as resolvins derived from EPA and DHA, and protectins and maresins derived from DHA [127,128,129]. EPA and DHA promote an anti-inflammatory environment by competing with arachidonic acid for the metabolic enzyme sites, thus reducing the production of pro-inflammatory eicosanoids. Indeed, clinical studies have shown that moderate- or high-dose supplementation with EPA and DHA ethyl esters increases n-3-PUFA-derived anti-inflammatory oxylipins, whilst reducing n-6-derived eicosanoids in human plasma [130,131,132]. Differences in the metabolism of EPA and DHA can introduce variability in their clinical effects. CYP450 isoforms have differences with respect to their specificity for each n-3-PUFA, whilst the metabolism of EPA and DHA may also be affected by differences in CYP450 catalytic activity due to gene polymorphisms [133,134]. Therefore, tissue-dependent, and inter-individual differences in the expression and activities of CYP450 isoforms may lead to variation in the anti-inflammatory to EPA and DHA.

n-3-PUFAs also inhibit pro-inflammatory signaling pathways. Activation of PPARγ by DHA and EPA leads to inhibition of the pro-inflammatory transcription factor nuclear factor (NF)-κB and up-regulation of adiponectin production by adipocytes [135,136,137]. n-3-PUFAs also promote anti-inflammatory effects by binding to the G-protein coupled receptor Free fatty acid receptor 4 (FFAR4/GPR120), which is highly expressed in adipose tissue and macrophages. Upon binding to n-3-PUFAs, FFAR4 promotes calcium-dependent regulation of physiological processes, including glucose metabolism and adipogenesis, and recruits and activates the scaffold protein β-arrestin-2 that has anti-inflammatory effects [138]. Specifically, the FFAR4/β-arrestin-2 complex inhibits NF-κB activation and the formation of the NOD-like receptor protein 3 (NLRP3) inflammasome in macrophages. NLRP3 is a cytosolic protein complex that senses metabolic stress signals, such as cholesterol crystals, triggering the activation of the inflammatory cytokines IL-1β and IL-18 [139,140].

A number of in vitro studies in different cell types demonstrate that EPA and DHA differentially regulate inflammatory mechanisms. In the monocytic cell line THP-1, EPA and DHA were shown to inhibit NF-κB activity through two distinct signaling cascades [141]. Moreover, a study in the human colon carcinoma cell line CaCo-2 demonstrated that both EPA and DHA alone can trigger FFAR4-dependent inhibition of NF-κB with the same efficacy but with different kinetics, possibly due to different affinity to the receptor [142]. On the other hand, experiments in HT29, another colon carcinoma cell, line show that EPA activates both calcium and β-arrestin-2-mediated signaling, whilst DHA showed no calcium activation and weak β-arrestin-2 recruitment [143]. Human FFAR4 shows two alternative splice variants; the short form triggers both calcium- and arrestin-mediated cascades and the long form only drives arrestin signaling [144]. The relative responses of EPA and DHA on FFAR4 may therefore depend on the relative expression of each isoform, which can be cell line-, tissue- and/or disease-specific, and the relative affinity of each fatty acid for each isoform. In a rat model of myocardial infarction, EPA was shown to have a stronger anti-inflammatory effect compared to DHA, by maintaining a higher PPARγ activity through inhibition of its phosphorylation [145].

4.2. Effects of EPA and DHA on Lipid Oxidation

Dyslipidemia and oxidative stress promote oxidized (ox)LDL and cholesterol crystal domain formation in endothelial cell membranes, which promote endothelial dysfunction and vascular inflammation. OxLDL triggers endothelial activation and the recruitment of macrophages leading to the formation of foam cells, which promote inflammation and atherosclerotic plague destabilisation [146]. Cholesterol crystal domains promote inflammation by activating the NLRP3 inflammasome, a protein complex that triggers caspase-induced activation of the cytokines IL-1β and IL-18 [147]. Both EPA and DHA, and their oxidation products, protect against LDL and cell membrane oxidation by directly scavenging reactive oxygen species, and by increasing endogenous antioxidant gene expression through activation of the cytoprotective transcription factor nuclear factor erythroid 2–related factor 2 (Nrf2) [148,149,150,151]. EPA was shown to have a stronger and more prolonged protective effect against LDL oxidation and cholesterol crystal domain formation, compared to DHA, in in vitro studies. This may be due to the chemical structure of EPA, and specifically the combination of its hydrocarbon length and location of double bonds, which allows more efficient electron stabilization [152,153].

4.3. Effects of EPA and DHA on Cell Death

Evidence supports the protective effects of n-3 PUFAs against apoptosis and cell death in atherosclerosis. EPA/DHA-rich oxLDL obtained from healthy subjects following fish oil supplementation, was shown to trigger less apoptosis in a monocytic cell line, compared to control oxLDL [154]. Furthermore, DHA supplementation (400–1600 mg/day) of healthy subjects for two weeks was shown to dose-dependently reduce the susceptibility of peripheral blood monocytes to oxLDL-dependent apoptosis [155]. In line with these clinical observations, in vitro studies report that oxidised EPA and DHA promote monocyte-derived macrophage apoptosis, but to a lesser extent than oxidised arachidonic acid, whilst DHA inhibits saturated fatty acid-induced endothelial cell apoptosis [156,157]. Maresin-1, a DHA metabolite, also inhibits endothelial cell inflammatory responses and apoptosis by activating PPARα-mediated signaling [158].

A study in an obese mouse model showed that an imbalance in saturated fatty acid/n-3 PUFA levels in peritoneal macrophage membranes is associated with impaired removal of apoptotic cells by efferocytosis, due to altered phosphatidylinositol 3-kinase activity. Intriguingly, the same study showed that fish oil supplementation normalised efferocytosis, indicating a role of n-3 PUFAs in effective clearance of apoptotic cells [159]. Indeed, DHA was shown to enhance the efferocytosis activity of macrophages by driving their differentiation to a more anti-inflammatory and pro-resolving phenotype via a PPARγ-dependent mechanism [160]. The DHA metabolite Resolvin D1 also promotes the efferocytosis of necroptotic cells by promoting cytoskeletal changes, and mitochondrial respiration and ATP production [161,162].

The above studies show the complexity of the molecular mechanisms underlying the effects of EPA and DHA on lipoprotein metabolism and oxidation, and inflammation. This is further compounded by the heterogeneity of the in vitro and in vivo models used, due to differences in experimental conditions, and tissue- and species-specific differences in signaling pathways. Furthermore, these models do not accurately represent the in vivo situation, where there is interaction between multiple cells, such as between immune cells with endothelial cells and adipocytes. Development of more reproducible and accurate models would enable a better understanding of mechanisms of action of EPA and DHA in order to determine their relative efficacies and optimise supplementation composition.

5. Epigenetic Determinants of n-3-PUFA Effects

An interesting observation from the ComparED study is that only 26% of subjects responded to either DHA or EPA. As the compliance of the study subjects was approximately 95%, the lack of response was probably not due to non-compliance but due to other factors [104]. This suggests that the variability observed in clinical studies could be a result of inter-individual differences in the response to n-3 PUFAs. These differences may stem from genomic differences, as a number of single nucleotide polymorphisms (SNPs) in genes involved in lipid metabolism and triglyceride synthesis have been associated with the effect of n-3 PUFAs on plasma TG [163,164,165,166]. Epigenetic mechanisms may also contribute to the variability in the response to n-3 PUFAs.

Availability of n-3 PUFAs, particularly DHA, is required for one-carbon metabolism that leads to the production of S-adenosyl methionine, the methyl donor of DNA methylation [167]. Indeed, a number of clinical studies in the last 10 years have demonstrated an effect of n-3 PUFAs on DNA methylation. DNA methylation analysis on blood from Yupik natives identified 27 differentially-methylated CpG sites that correlated with the levels of n-3 PUFA intake, including the fatty acid synthase gene, and the aryl-hydrocarbon receptor repressor gene involved in antioxidant protection [168]. Furthermore, a study in lactating infants and adult men reported a strong association between EPA and arachidonic acid levels with global DNA methylation, with EPA having a stronger association than arachidonic acid [169]. A randomised control study in healthy or chronic kidney disease patients showed that 4 g/day n-3 PUFAs (1.8 g EPA, 0.2 g DPA, 1.5 g DHA ethyl esters) for 8 weeks, led to altered CpG methylation in the 5′-regulatory region of fatty acid desaturase (FADS) 2 gene, encoding Δ6 desaturase, and elongase (ELOVL)-5 that encodes elongase 5, in peripheral blood mononuclear cells (PBMCs) [170]. A study conducted in PBMCs from obese subjects on a low-calorie diet found that supplementation with n-3 PUFA-rich fish oil for 8 weeks altered CpG methylation at a specific site in the CD36 gene, a membrane glycoprotein that is involved in lipid uptake, including oxLDL internalisation, in macrophages. Nonetheless, the changes in methylation reported were very small suggesting that they may not affect gene expression [171]. Studies have also investigated the effects of prenatal n-3 PUFA supplementation on the infant epigenome.

Supplementation of pregnant women with DHA (400 mg/day) from gestation week 18–22 until birth was associated with increased methylation of the imprinted genes insulin-like growth factor 2 (IGF2) and H19 in umbilical cord mononuclear cells, compared to the control group [172]. IGF2/H19 differentially methylated regions have been associated with the risk of higher birth and early life weight [173,174]. A subsequent randomised control study reported that supplementation of pregnant women with DHA-rich fish oil (800 mg/day) from gestational week 20 until delivery did not significantly alter global DNA methylation in blood leukocytes, but induced small changes in the methylation of a subset of genes involved in lipid metabolism, appetite regulation and immune function [175]. Similarly, a study on 9-month-old infants supplemented with EPA/DHA-rich fish oil (1.6 g/day) for 9 months, showed no statistically-significant difference in global DNA methylation [176].

The above studies show that n-3 PUFAs may promote their clinical effects through epigenomic re-programming. However, the findings show small changes in DNA methylation, which may not have important functional consequences. This may be due to the small sample size of the studies or to insufficient n-3 PUFAs dose. Furthermore, as DNA methylation patterns are tissue-specific, epigenetic analysis of other tissues may lead to a more complete picture of the epigenetic effects of n-3 PUFAs. Importantly, functional studies are required in order to understand the clinical implications of these changes.

6. Conclusions and Outlook

CVD is the leading cause of mortality worldwide [68]. It is more than clear nowadays that n-3 PUFAs do not only serve as an inert form of energy storage [177]. Epidemiological studies and randomized control trials have reported that n-3 PUFAs might reduce cardiovascular events [54]. The evidence of the benefits of n-3 PUFAs is stronger in secondary than in primary prevention settings [34]. Several potential mechanisms for the cardioprotective effect of n-3 PUFAs have been proposed, with the effect on blood lipids being the most scientifically proven [44,45]. Particularly, their most consistent effect is the reduction in serum TG [54]. Nonetheless, there are conflicting findings regarding their clinical benefits, leading to uncertainty regarding their use for preventing atherosclerosis [26]. Several factors need to be considered when designing clinical studies, especially when natural ingredients and natural extracts are tested for efficacy. Firstly, as the TG-lowering effects of n-3 PUFA are dose-dependent, the use of consistent, and preferably high doses, across studies may lead to more meaningful results [102]. Secondly, the selection of suitable study subjects is crucial as n-3 PUFAs were shown to have a more pronounced effect in patients with hypertriglyceridaemia [61]. Thirdly, the selection of the chemical form, composition, and EPA/DHA ratio of n-3 PUFA is an important factor, as it affects their overall molecular effects. Another important factor in accurately collecting data is the consideration of the run-in period that is about 3 to 6 months from the time of the supplement first consumed. A period that is known for some natural products as required to indicate activity.

Optimising the EPA and DHA content of n-3 PUFA supplements requires a clear understanding of the molecular mechanisms driving their effects. Currently most of the available mechanistic data have been obtained from in vitro and animal models, whilst very limited data from humans are available [178]. Moving towards 3-dimentional organoid cultures of patient-derived cells would permit studying the effects of n-3 PUFAs in an environment that more closely resembles the human in vivo situation. This approach will generate data that can be more easily translated into clinical studies.

Another important issue is identifying or predicting which patients respond to n-3 PUFA supplementation in order to provide more personalised management. More extensive and better-powered epigenomic studies, accompanied by functional studies, would elucidate the role of epigenetic mechanisms in the cellular effects of n-3 PUFAs, as well as epigenetic biomarkers for predicting responders or non-responders. In line with this notion, transcriptomics analysis of PBMCs of responders or non-responders to the TG-lowering effect of EPA/DHA-rich fish oil (1.6 g/day), identified differentially regulated genes between the two groups [179]. Future studies could use transcriptomics and metabolomics analysis to identify gene and metabolite signatures, which could be used as biomarkers for predicting patients that would benefit from the use of n-3 PUFAs. This could involve performing DNA methylation, transcriptomics and/or metabolomics analysis of blood plasma samples before and after treatment with n-3 PUFAs supplements. The potential biomarkers could then be correlated with blood lipids, and inflammatory and thrombotic markers, and their predictive accuracy determined. Furthermore, integration of transcriptomics and metabolomics data can be used in order to identify new n-3-PUFA-regulated pathways, which could lead to a better understanding of the molecular mechanisms underlying their effects.

Concluding, the existing clinical trials are not appropriately designed considering all important parameters to correctly evaluate their efficacy in the treatment of different CVDs. High doses and ratio with DHA in excess might be the answer to the mystery. All the rational and biochemical networks that can be affected by the use of these natural molecules as supplements suggest the need for carefully designed basic experimental studies and clinical trials in order to reach accurate conclusions on their efficacy in treating different CVDs.