- © 2014 American Society for Nutrition
Fish-oil supplementation alters numbers of circulating endothelial progenitor cells and microparticles independently of eNOS genotype1,2,3,4
Abstract
Background: Emerging cellular markers of endothelial damage and repair include endothelial microparticles (EMPs) and endothelial progenitor cells (EPCs), respectively. Effects of long-chain (LC) n−3 (omega-3) polyunsaturated fatty acids (PUFAs) and the influence of genetic background on these markers are not known.
Objective: We investigated effects of fish-oil supplementation on both classical and novel markers of endothelial function in subjects prospectively genotyped for the Asp298 endothelial nitric oxide synthase (eNOS) polymorphism and at moderate risk of cardiovascular disease (CVD).
Design: A total of 84 subjects with moderate risk of CVD (GG: n = 40; GT/TT: n = 44) completed a randomized, double-blind, placebo-controlled, 8-wk crossover trial of fish-oil supplementation that provided 1.5 g LC n−3 PUFAs/d. Effects of genotype and fish-oil supplementation on the blood lipid profile, inflammatory markers, vascular function (by using peripheral artery tonometry), and numbers of circulating EPCs and EMPs (by using flow cytometry) were assessed.
Results: There was no significant effect of fish-oil supplementation on blood pressure, plasma lipids, or plasma glucose, although there was a trend (P = 0.069) toward a decrease in the plasma triglyceride concentration after fish-oil supplementation compared with placebo treatment. GT/TT subjects tended to have higher concentrations of total cholesterol and low-density lipoprotein cholesterol, but vascular function was not affected by either treatment or eNOS genotype. Biochemical markers of endothelial function were also unaffected by treatment and eNOS genotype. In contrast, there was a significant effect of fish-oil supplementation on cellular markers of endothelial function. Fish-oil supplementation increased numbers of EPCs and reduced numbers of EMPs relative to those with placebo treatment, which potentially favored the maintenance of endothelial integrity. There was no influence of genotype for any cellular markers of endothelial function, which indicated that effects of fish-oil supplementation were independent of eNOS genotype.
Conclusion: Emerging cellular markers of endothelial damage, integrity, and repair appear to be sensitive to potentially beneficial modification by dietary n−3 PUFAs. This trial was registered at www.controlled-trials.com/isrctn as ISRCTN76272133.
INTRODUCTION
Endothelial dysfunction (ED)5 is associated with classical and emerging risk factors of cardiovascular disease (CVD), correlates with disease progression, and is an independent predictor of risk of future cardiovascular events (1). The assessment of ED, either functionally by endothelium-dependent vasodilation or by circulating endothelial biomarkers, is, therefore, of potential clinical value. Circulating endothelial biomarkers are typically derived from molecular pathways related to aspects of endothelial function and include direct products of endothelial cells such as measures of nitric oxide (NO) bioavailability, adhesion molecules, inflammatory cytokines, and regulators of thrombosis (1). However, the recognition that endothelial dysfunction reflects a disrupted balance between endothelial injury and repair has led to the identification of endothelial progenitor cells (EPCs) and endothelial microparticles (EMPs) as noninvasive cellular makers of endothelial function (2, 3). EPCs are small (<15 μm), immature precursor cells detectable in plasma and bone marrow, whereas EMPs are small vesicles (0.1–1 μm) released from activated, damaged, or apoptotic endothelial cells. Numbers of EPCs are thought to reflect the body's capacity for endothelial repair, whereas EMPs are directly indicative of endothelial stress and damage. The addition of EMP numbers to the Framingham risk score model improves the prediction power of future cardiovascular events, and EMPs are elevated in individuals with metabolic syndrome, hypertension, and obesity (4). In contrast, reduced EPC numbers and function are associated with CV risk factors (5, 6), and there is interest in the potential role of EPCs as prognostic and diagnostic markers of CVD (6).
Pharmacologic and lifestyle interventions, including diet, have the potential to modify vascular and endothelial function, providing a basis for primary prevention. There has been moderate evidence that SFAs adversely affect vascular function (7), whereas n−3 PUFAs improve endothelial function assessed either functionally or by classical biomarkers (8, 9). However, there are little or no data on the effects of n−3 PUFAs on emerging cellular markers of CVD, including EPCs and microparticles from all sources (that include endothelial), and the genetic background may be a confounder. NO, which is produced by endothelial nitric oxide synthase (eNOS), plays a central role in maintaining endothelial function and a common variant of the eNOS gene located in exon 7 (G894>T), which modifies its coding sequence (Glu298>Asp), has been linked to CVD (10). In addition, beneficial effects of n−3 PUFA supplementation on postprandial flow-mediate dilatation (FMD) (11) and markers of CVD (12) appear to be more pronounced in Asp298 carriers. Therefore, we hypothesized that dietary n−3 PUFA supplementation would have greater impacts on both classical and novel markers of endothelial function in Asp298 carriers. In this randomized, double-blind, placebo-controlled, crossover trial, we investigated the effects of fish-oil supplementation on markers of endothelial function in individuals at moderate risk of CVD who were prospectively recruited for eNOS genotype.
SUBJECTS AND METHODS
Ethics and trial registration
The study protocol was reviewed and approved by the University of Reading Research Ethics Committee (project 09/69). The trial was registered with Current Controlled Trials, London (www.controlled-trials.com/isrctn; ISRCTN76272133) and conducted according to the guidelines of the Declaration of Helsinki.
Participants
From October 2010 to April 2012, subjects aged 21–65 y were recruited from the population in and around Reading, United Kingdom, through newspaper and poster advertisements, e-mail, and radio. Persons interested in participating in the study completed a health and lifestyle questionnaire by telephone or e-mail. Candidates who met inclusion criteria underwent anthropometric measurements and biochemical tests in a screening visit to identify individuals at mild-to-moderate risk of CVD (RR >1.5) on the basis of scoring ≥2 points in one or more risk factors from a scoring tool adapted from the Framingham risk-score system (Table 1) (13–15). Potential candidates for the study were advised that they would be expected to maintain a low consumption of n−3 (omega-3) fatty acids during the study, refrain from the use of all supplements, and maintain their body weights. Exclusion criteria included blood pressure >160/100 mm Hg; hyperlipidemia (total cholesterol concentration >8 mmol/L); anemia (hemoglobin concentration <12.5 g/L in men and <11.5 g/L in women); BMI (in kg/m2) <18.5 or >38; diabetes (diagnosed or fasting glucose concentration >7 mmol/L) or other endocrine disorders; angina, stroke, or any vascular disease in the past 12 mo; renal or bowel disease or a history of choleostatic liver or pancreatitis; the use of medication for hyperlipidemia, hypertension, hypercoagulation, inflammation, or depression; pregnancy or breastfeeding; planning to start or on a weight-reducing regimen; intense aerobic exercise (>3 × 20 min/wk); smoking; taking dietary supplements; consuming large amounts of oily fish; alcohol misuse or intakes >21 units/wk for men and >15 units/wk for women (one UK unit was defined as 10 mL or 8 g pure alcohol). All subjects gave written informed consent before participation in the study.
Scoring tool for recruitment1
Sample size
Studies that reported significant effects of fish oil on endothelial function assessed by using FMD typically involve 12–30 subjects/group although doses and durations have varies (9). However, 34 subjects/group are required to detect a significant effect of the Asp (T) allele on FMD with a 2-sided significance level of 5% and a power of 80% (16). Therefore, we aimed to recruit 40 subjects in each group to allow for a 15% dropout. A total of 171 subjects were screened. Of these, 91 subjects were eligible to participate in the trial and were genotyped for the Glu298Asp (G/T) polymorphism and assigned to a GG or GT/TT group. Because the TT genotype is relatively rare (∼10%), homozygotes (TT) and heterozygotes (GT) were grouped together.
Study design
The study was a double-blind, placebo-controlled, crossover trial that consisted of 2 treatment periods, each with a duration of 8 wk, separated by an 8-wk washout interval. Within each genotype group, participants were stratified according to sex, BMI, and score and were randomly allocated to receive either a placebo (control) in the first treatment period followed by fish oil in the second treatment period or the reverse; random assignment of subjects for treatment order was performed with Minim software (version 1.5; London Hospital Medical College) (17). Vascular measurements and blood sampling were conducted at the beginning and end of each treatment period after an overnight fast. Participants were asked to refrain from intensive exercise and alcohol consumption 24 h before study visits and to continue with their normal diet consuming no more than one portion of oily fish per fortnight. At baseline, participants were also asked to complete a food-frequency questionnaire to characterize the background habitual diet. Participants were asked to consume 3 capsules fish oil/d (EPAX 6000 TG; EPAX), each of which contained 300 mg EPA and 200 mg DHA as triacylglycerol (providing a total daily dose of 0.9 g EPA plus 0.6 g DHA) or placebo capsules of identical appearance that contained corn oil (predominantly comprised of 54% linoleic, 29% oleic, and 12% palmitic acids) (EPAX) for 8 wk.
Anthropometric measures, blood pressure, and peripheral artery tonometry
Participants emptied pockets and removed all footwear and heavy clothing before anthropometric measurements were taken. Height was measured to the nearest 0.5 cm by using a wall-mounted calibrated stadiometer. Weight was measured by using a Tanita BC-418 digital scale (Tanita Europe BV) to the nearest 0.1 kg by using standard settings (normal body type and −1 kg for clothing). Waist circumference was measured to the nearest 0.5 cm at a horizontal plane midway between the lowest rib and iliac crest. Blood pressure was measured in triplicate by using a validated semiautomatic sphygmomanometer after a minimum of 5 min rest in the seated position.
Peripheral artery tonometry was assessed by using an EndoPAT 2000 device (Itamar Medical Ltd) according to the manufacturer's guidelines. Briefly, participants were in a supine position for a minimum of 20 min before measurements were taken in a quiet, dimmed, and temperature-controlled (21–24°C) room. Subjects were asked to remain as still and silent as possible during the entire measurement period. The occlusion cuff was placed above the elbow on the nondominant arm (test arm), and fingertip plethysmography probes were placed on index fingers of each hand. Pulse wave amplitude (PWA) and heart rate were continuously recorded in both fingers throughout the 15-min testing. Recordings involved a 5-min baseline period followed by a 5-min occlusion period, during which the cuff was inflated at 60 mm Hg above the systolic blood pressure (minimally 200 mm Hg and maximally 300 mm Hg), and a 5-min postocclusion period after cuff release. From PWA recordings, the EndoPAT software automatically calculated the reactive hyperemia index (RHI) through a computer algorithm as the ratio of the mean PWA during hyperemia (90–150 s of the postocclusion period) to the mean PWA during baseline in the test arm divided by the corresponding ratio in the control arm and multiplied by a baseline correction factor. Higher values of RHI correlate with better endothelial function with RHI >1.67 considered normal. The EndoPAT device also generated an augmentation index (AI) and augmentation index adjusted for a heart rate of 75 beats/min (AI@75), which served as measures of arterial stiffness.
Blood sample collection and assays
After EndoPAT measurements, blood samples were collected into sodium citrate tubes (for the analysis of von Willebrand factor, EPCs, and EMPs), serum separating tubes (for the analysis of blood lipids, C-reactive protein, and glucose), sodium heparin tubes (for the analysis of soluble cellular adhesion molecules and NO), and K3EDTA-coated tubes (for the analysis of the plasma phospholipid fatty acid composition). After blood sample collection, sodium citrate and serum separating tubes were stored at room temperature for 15 min, whereas the other tubes were kept on ice. With the exception of samples for the analysis of EPCs and EMPs, blood samples were centrifuged at 1800 × g for 15 min at room temperature or 4°C, as appropriate, ≤30 min of withdrawal. Plasma and serum samples were divided into aliquots and stored at −20°C until analysis.
DNA extraction and eNOS genotyping
DNA was isolated from the buffy coat layer of 10 mL screening blood drawn into K3EDTA-coated tubes with the use of a QIAmp DNA Blood Mini Kit (Qiagen Ltd). After DNA isolation, allelic discrimination of the Glu/Glu (GG), Glu/Asp (GT), and Asp/Asp (TT) eNOS gene variants was conducted by using TaqMan polymerase chain reaction technology (Applied Biosystems 7300 Instrument; Applied Biosystems) and Assay-on-Demand single nucleotide polymorphism genotyping assays (Applied Biosystems).
Plasma lipids and plasma glucose
Plasma lipid (triacylglycerol, total cholesterol, and HDL cholesterol) and glucose concentrations were determined with an ILAB 600 clinical chemistry analyzer (Instrumentation Laboratory Ltd) by using enzyme-based colorimetric kits (Instrumentation Laboratory Ltd) with the inclusion of appropriate controls. The LDL-cholesterol concentration was estimated by using Friedewald's formula. Intraassay and interassay CVs were <2%.
Inflammatory markers
ELISAs were used for the measurement of soluble intercellular adhesion molecule 1, soluble vascular cell adhesion molecule, E-selectin (R&D systems), and von Willebrand factor antigen (Abnova) according to the manufacturers’ instructions. Mean intraassay and interassays CVs were <11.1% and <15.3%, respectively.
Plasma NO
Plasma samples were analyzed for nitrite and nitrate by using ozone chemiluminescence. In brief, total nitrite and nitrate concentrations were determined by the addition of 20 μL plasma samples to 0.1 mol/L vanadium (II) chloride in 1 mol/L hydrochloric acid refluxing at 80°C. These conditions caused the reduction of nitrate, nitrite, nitrosothiols, nitrosamines, iron nitrosylhemoglobin, and nitrosohemoglobin to NO, which was quantified by chemiluminescence (Model 88 et CLD; Eco Physics).
EPCs
Venous blood was collected into 4.5 mL sodium citrate evacuated tubes and processed within 3 h. For EPC analysis, 20 μL allophycocyanin-conjugated antihuman CD34 monoclonal antibody (mAb) and 20 μL phycoerythrin-conjugated antihuman kinase insert domain receptor (KDR) mAb were added to 200 μL whole blood in a TruCount tube (BD Biosciences). Fluorescence-minus-one and isotype controls with isotype-identical antibodies were performed for each sample (all mAb obtained from BD Biosciences). Sample tubes were incubated for 20 min at room temperature in the dark. Red blood cells were lysed in 2 steps by adding 1 mL 1× FACS lysing solution (BD Biosciences), mixing briefly before adding an additional 760 μL 1× FACS lysing solution (BD Biosciences) to each tube, and followed by vortexing and incubating for 30 min at room temperature in the dark. A sample acquisition was performed for a minimum of 500,000 total events on a FACSCanto II flow cytometer (BD Biosciences) with FACSDiva 6.1.2 software (BD Biosciences). After appropriate gating (Figure 1), the number of CD34+KDR+ cells were defined as EPCs and expressed as the number of cells per milliliter of blood.
Gating strategy for EPC enumeration. A: A live gate in a scatter plot was used to exclude debris but not beads. A total of 500,000 events were acquired outside this live gate to allow the analysis of ≥100 EPCs/sample. B: A gate was set around cells positive for CD34 with low side scatter. C: Back gating was performed to confirm the cell size and clustering of EPCs in the lymphocyte region. Clustered events were gated to remove residual debris from either side of the EPC population, including circulating endothelial cells, which are larger. D: Cells double positive for CD34 and KDR in this gate were defined as EPCs. E: The cutoff for KDR+ events was determined by using a fluorescence-minus-one control. A, area; APC, allophycocyanin; EPC, endothelial progenitor cells; FSC, forward scatter; Ig2G1, immunoglobulin G1; KDR, kinase insert domain receptor; PE, phycoerythrin; SSC, side scatter.
EMPs and platelet microparticles
Venous blood was collected into a 4.5-mL sodium citrate evacuated tube, the first tube was discarded, and the remainder was processed ≤1 h of collection. Samples were centrifuged for 10 min at 160 × g to prepare platelet-rich plasma, and 250 μL platelet-rich plasma was further centrifuged for 6 min at 1000 × g to obtain platelet-poor plasma. Platelet-poor plasma (50 μL) was stained with 20 μL CD42b conjugated to fluorescein isothiocyanate and 20 μL of CD31 conjugated to phycoerythrin in a TruCount tube (BD Biosciences). Fluorescence-minus-one and isotype controls with isotype-identical antibodies were performed for each sample (all mAbs were obtained from BD Biosciences). Samples were incubated for 20 min at room temperature in the dark, diluted with 910 μL 0.2-μmol/L filtered phosphate-buffered saline, and analyzed on a BD FACSCalibur flow cytometer (BD Biosciences) by using a medium flow-rate setting. A microparticle phenotype analysis was based on size and fluorescence. Light scatter and fluorescence channels were set at logarithmic gain. A mix of fluorescent beads (Biocytex) of varied diameters (0.5, 0.9, and 3 μm) was used for the size calibration, and 500,000 events <1 μm in size were acquired. From these events, EMPs were defined as CD31+CD42b− particles, and platelet microparticles (PMPs) were defined as CD31+CD42b+ particles (Figure 2). Values were reported as counts per microliter of blood.
Gating strategy for EMP and PMP enumeration. Light scatter and fluorescence channels were set at logarithmic gain. A: Calibration of fluorescent beads of 3 different sizes (0.5, 0.9, and 3 μm). B: Upper detection limit of the microparticle gate on a scatter plot was established with the position of 0.9-μm beads, whereas the lower side was defined by the threshold. A total of 500,000 events were acquired in the microparticle gate. C: Within the microparticle gate, events positive for CD31 were identified. D: EMPs were defined as CD31+CD42b−, whereas PMPs were defined as CD31+CD42b+. E: Cutoff for CD42b+ was determined by using a fluorescence-minus-one control. EMP, endothelial microparticle; FITC, fluorescein isothiocyanate; FL, filter; FSC, forward scatter; IgG1, immunoglobulin G1; PE, phycoerythrin; PMP, platelet microparticle; SSC, side scatter.
Plasma phospholipid fatty acid–composition analysis
Total lipid was extracted with chloroform:methanol (2:1 vol:vol) that contained butylated hydroxytoluene as an antioxidant. Phospholipid fatty acids were isolated by using solid-phase extraction on aminopropylsilica cartridges (Varian). Fatty acid methyl esters (FAMEs) were generated by reaction with methanolic sulphuric acid (2% vol:vol H2SO4) at 70°C for 1 h. FAMEs were analyzed by using gas chromatography on an HP-6890 Series GC System (Hewlett-Packard) equipped with a flame-ionization detector and a CP-Sil 88 fused silica capillary column (50 m × 0.25-mm inside diameter × 0.20-μm film thickness; Varian). Helium at 24.0 cm/s was used as the carrier gas, and the split-splitless injector was used with a split ratio of 10:1. Injector and detector temperatures were 240°C and 260°C, respectively. The column oven temperature was maintained at 130°C for 3 min after sample injection and was programmed to increase at 4°C/min to 240°C and held at this temperature for 6 min (total run time: 36.5 min). FAMEs were identified by comparison with retention times against known standards (Supelco FAME Mix C4-C24, PUFA-3 Menhaden Oil Standards; Supelco) and quantified by using HP-Chemstation software (version A10.01; Hewlett-Packard) for gas-chromotography systems. Results were expressed as the relative percentage of total phospholipid fatty acids.
Statistical analysis
Statistical analyses were performed with SPSS software (version 20; IBM SPSS Statistics). Before analysis, the data distribution and normality was checked visually and by using the Kolmogorov-Smirnov and Shapiro-Wilk tests, and the ln transformation was used for skewed variables. Differences in baseline characteristics between eNOS-genotype groups were tested by using the t test and chi-square test for continuous and categorical variables, respectively. To determine independent and interactive effects of genotype and treatment, mean changes (from baseline to the end of the 8-wk treatment) were compared by using a linear mixed model. Change-from-baseline values for each variable and participant were calculated by subtracting week 0 from week 8 values and week 24 from week 16 values. These data were entered into a linear mixed model, which included the sequence, treatment, and period. Fixed effects always included in the final model were baseline values of the assessed variable, sequence, period, treatment, genotype, age, sex, and BMI. The treatment × genotype interaction was initially included in the model, but in the absence of any significant interactions, it was removed in the final analysis of main effects. Fixed-effect covariates were retained in the model regardless of their degrees of significance. The subject was included as a random factor within the linear mixed model. Residual analyses were conducted to check that assumptions of the model were justifiable. There were no effects of the period or sequence in the model for any outcome. When significant interactions were shown, slice tests were used to test for effects within levels of an interaction. Results are presented as means (±SEMs). P < 0.05 was considered significant.
RESULTS
Of 91 subjects who participated in the study, 84 subjects (n = 40 with GG and n = 44 with GT/TT) completed the 2 intervention periods and all 4 visits. Baseline characteristics and habitual dietary intakes of subjects according to eNOS genotype are shown in Tables 2 and 3, respectively. The final study population was middle aged (age: 47.6 ± 1.3), normotensive (systolic blood pressure/diastolic blood pressure: 123.2 ± 1.5/76.0 ± 1.0 mm Hg), slightly overweight (BMI: 26.1 ± 0.5), and mildly hypercholesterolemic (5.32 ± 0.10 mmol/L), with no significant differences between genotype groups at baseline for subject characteristics. Of 44 subjects in the GT/TT group, 7 subjects were homozygous for TT, and the remaining 37 subjects were GT.
Baseline characteristics of subjects completing the study according to eNOS genotype1
Baseline dietary intake of subjects according to eNOS genotype1
Compliance
Compliance was monitored by capsule counts and changes in the plasma fatty acid composition. Subjects were provided with capsules in excess of requirements, and remaining capsules at the end of the 8-wk treatment period were counted. The average compliance was 93% and did not vary according to the treatment (control compared with fish oil: 92% compared with 94%, respectively; P = 0.175) or genotype groups (GG compared with GT/TT: 91% compared with 94%, respectively; P = 0.081). Good compliance by capsule counts was reflected in a significant increase in plasma phospholipid EPA and DHA after fish-oil supplementation (2.70 ± 0.14% and 2.06 ± 0.12%, respectively) than with the placebo (−0.05 ± 0.10% and 0.05 ± 0.09%, respectively).
Effect of fish-oil supplementation and eNOS genotype on blood pressure and biochemical variables
There was no significant effect of fish-oil supplementation on blood pressure, plasma lipids, or plasma glucose, although there was a trend (P = 0.069) toward a decrease in the plasma triacylglycerol concentration after fish-oil supplementation than with the placebo in the group as a whole (Table 4). In addition, there was a modest association between the reduction in plasma triacylglycerol concentrations and increase in plasma phospholipid concentrations of EPA (r = −0.19, P = 0.017) and DHA (r = −0.19, P = 0.016). An overall genotype effect was evident for total cholesterol and LDL cholesterol, with GT/TT subjects tending to have higher concentrations. No significant diet × genotype interactions were observed for any of the previously detailed measures.
Effect of FO supplementation and eNOS genotype on blood pressure and biochemical variables1
Effect of fish-oil supplementation and eNOS genotype on markers of vascular function
Endothelial function, as assessed by using the RHI or Framingham RHI, was not affected by the treatment or eNOS genotype (Table 5). There were also no overall treatment or genotype effects on arterial stiffness (AI or AI@75) or biochemical markers of endothelial function. In contrast, there was a significant effect of fish-oil supplementation on cellular markers of endothelial function. EPC numbers increased in response to fish-oil supplementation compared with the placebo (126.5 ± 37.8 compared with 5.17 ± 36.7, respectively; P = 0.020) (Figure 3). In addition, fish-oil supplementation reduced EMP numbers relative to those with the placebo (−8.75 ± 1.42 compared with −2.74 ± 1.38, respectively; P = 0.001) but had no significant effect on PMP numbers (−35.0 ± 19.7 compared with −26.7 ± 16.8, respectively; P = 0.217) (Figure 4). There was no influence of genotype for any of the cellular markers of endothelial function (P = 0.361, P = 0.369, and P = 0.285 for EPCs, EMPs, and PMPs, respectively) and no significant interactions between treatment and genotype, which indicated that effects of fish-oil supplementation were independent of eNOS genotype.
Effect of FO supplementation and the eNOS genotype on vascular function variables1
Effect of fish-oil supplementation and eNOS genotype on numbers of circulating EPCs. Data are adjusted mean (±SEM) changes from baseline in EPC numbers after 8 wk of treatment with fish oil (black bars) or placebo (white bars) according to eNOS genotype (GG, n = 40; GT/TT, n = 44). A significant treatment effect (P = 0.017) was observed when data were analyzed with the use of a general linear mixed model with EPC at baseline, sequence, period, treatment, treatment × genotype, age, sex, and BMI as fixed effects and subjects as random effects. There was no significant effect of genotype (P = 0.363) or treatment × genotype interactions (P = 0.375). EPC, endothelial progenitor cell.
Effect of fish-oil supplementation and eNOS genotype on numbers of circulating EMPs and PMPs. Data are adjusted mean (±SEM) changes from baseline after 8 wk of treatment with fish oil (black bars) or placebo (white bars) according to eNOS genotype (GG, n = 40; GT/TT, n = 44). A: For EMPs, a significant treatment effect (P = 0.001) was observed when data were analyzed with the use of a general linear mixed model with EMP at baseline, sequence, period, treatment, treatment × genotype, age, sex, and BMI as fixed effects and subjects as random effects. There was no significant effect of genotype (P = 0.361) or treatment × genotype interactions (P = 0.155). B: For PMPs, there was no significant effect of treatment (P = 0.218) or genotype (P = 0.260), and there was no treatment × genotype interaction (P = 0.285). EMP, endothelial microparticle; PMP, platelet microparticle.
DISCUSSION
The impact of fish-oil containing long-chain (LC) n−3 PUFAs on the secondary prevention of CVD is currently controversial (19–21). LC n−3 PUFAs have been reported to improve risk factors for CVD, including plasma triacylglycerol (22), heart rate (23), blood pressure (24), and vascular function (24). The Glu298>Asp polymorphism in the eNOS gene is linked to CVD (10) and reduced endothelial function (25), and there has been some evidence to suggest interactions with LC n−3 PUFAs (11, 12). The current study showed, for the first time to our knowledge, that 8 wk supplementation with fish oil (containing 1.5 g LC n−3 PUFAs/d ) significantly increased numbers of circulating EPCs and decreased numbers of EMPs, which suggested an improved endothelial maintenance and repair and reduced endothelial damage. Improvements in cellular markers of endothelial function were not accompanied by improvements in vascular function as assessed by the EndoPAT device or changes in biochemical markers and were not influenced by eNOS genotype.
There is growing evidence for effects of LC n−3 PUFAs on vascular function. However, in the current study, neither endothelial function, as measured by using the RHI or Framingham RHI, nor arterial stiffness, as measured by using the AI or AI@75, were affected by LC n−3 PUFA supplementation. Our findings agree with a recent study in middle-aged healthy adults supplemented with 0.85 or 3.4 g LC n−3 PUFAs/d (26). The influence of eNOS genotype on vascular function is not clear, although it has been shown that TT, but not GT, individuals have a significantly lower FMD than do GG carriers (16), and the postprandial enhancement of FMD after LC n−3 PUFA supplementation is greater in TT than GG individuals (11). Because of the low prevalence (∼10%) of homozygotes for the T allele, TT and GT individuals are frequently grouped together, as they were in the current study, and this grouping may reduce the ability to identify a genotypic influence.
Despite the lack of significant effects of LC n−3 PUFAs on vascular function or biochemical markers, there was a significant effect of LC n−3 PUFAs on numbers of circulating EPCs and EMPs. Recent in vitro and animal studies have suggested that EPA and DHA improve EPC numbers and functionality (27, 28). To our knowledge, only one human study has investigated effects of fish-oil supplementation on EPC numbers. In contrast to our findings, Wong et al (29) reported that 12 wk supplementation with fish oil at a dose of 4 g/d in type 2 diabetes patients had no effect on EPC counts). However, in the study, EPCs were classified as CD133+KDR+ rather than CD34+KDR+; the CD133+KDR+ population is even more rare than the CD34+KDR+ population is, and CD133 expression appears to be restricted to the most-primitive EPCs and lost during maturation (30). Furthermore, it is the CD34+KDR+ phenotype that has been shown to be a biomarker of CVD and is, therefore, arguably more relevant (4–6). The beneficial effects of LC n−3 PUFAs on EPC numbers were independent of eNOS genotype, which was consistent with the null relation between EPCs and eNOS genotype, including the Glu298Asp polymorphism, in renal transplant patients (31).
The mechanisms that underlies the effects of LC n−3 PUFAs on EPC numbers are unclear. However, recent evidence suggests that NO is an essential mediator of EPC mobilization from bone marrow (32). Thus, the upregulation in NO bioavailability in the bone marrow could be responsible for the increase in EPC numbers, although in the current study, there was no effect of LC n−3 PUFA supplementation on circulating NO concentrations. An alternative mechanism is that the LC n−3 PUFA–induced modification of lipid rafts could play a role in altered cell migration (33), or the incorporation of LC n−3 PUFAs into membranes could regulate the expression and activity of NADPH oxidase to influence EPC survival rates (34). Greater survival of EPCs has also been shown after treatment with peroxisome proliferator–activated receptor γ (PPAR-γ) agonists, which increase numbers of EPCs in blood and bone marrow by the prevention of apoptosis via the phosphatidylinositol-3-kinase/Akt signaling pathway (35, 36). Because LC n−3 PUFAs have been suggested to be PPAR-γ agonists, this represents an additional potential mechanism. The increase in circulating EPC could also be explained by the mobilization or enhanced survival through a stromal cell derived factor 1–mediated mechanism (37).
In addition to potentially improving the endothelial capacity for repair, the reduction in circulating numbers of EMPs in the current study also suggested a decreased endothelial damage as a result of LC n−3 PUFA supplementation. Although the precise mechanisms that underlie EMP release in vivo are unclear, it is possible that the remodeling of endothelial cell membranes by LC n−3 PUFAs reduces cell activation and/or apoptosis, resulting in reduced EMP release, but a change in EMP numbers could also reflect increased clearance. Notably, oxidative stress (38) and PPAR-γ ligands (39) modulate numbers of circulating EMPs. One other study investigated the effects of fish-oil supplementation on EMP numbers; Del Turco et al (40) reported that supplementation with 4.3 g LC n−3 PUFAs/d of for 12 wk in patients with a previous myocardial infarction had no significant effect on numbers of total microparticles or EMPs than did an olive oil placebo, although their study was relatively small. Del Turco et al (40) also reported lower numbers of platelet-derived and monocyte-derived microparticles, whereas in the current study, there was no effect of fish oil on numbers of PMPs; reasons for differential effects on microparticle subtypes are unclear, although historically, effects of fish oil on platelet function have tended to be relatively modest (41).
One of the hallmarks of high-dose supplementation with LC n−3 PUFAs is the lowering of blood triacylglycerol concentrations (22). Effects of LC n−3 PUFAs on blood triacylglycerol concentrations have been consistently reported at doses >2 g/d; below this dose, results have been variable (22). The trend for triacylglycerol lowering after supplementation with 1.5 g LC n−3 PUFAs/d (albeit NS) in the current study was, therefore consistent, with the literature. Note that several of the large intervention studies that examined effects of fish oil on cardiovascular outcomes reported significant beneficial effects of fish oil in the absence of significant triacylglycerol lowering (42, 43), suggesting that the impact of LC n−3 PUFAs in CVD involves factors other than changes in blood lipids. Also pertinent is the fact that effects of fish-oil supplementation on plasma triglyceride concentrations tend to be more significant in individuals with raised baseline concentrations (22), and in the current study, concentrations were in the normal range.
Inflammatory factors, adhesion molecules, and endothelial markers are also potential targets for the influence of LC n−3 PUFAs. However, recent trials have been inconsistent with respect to effects of LC n−3 PUFAs on serum concentrations of inflammatory markers in healthy individuals (44) and with respect to the influence of eNOS genotype (45, 46). In the current study, there was no effect of fish-oil supplementation on inflammatory markers, and there was no influence of eNOS genotype. Circulating total nitrite and nitrate concentrations were not affected by supplementation with fish oil, and there was a lack of an association between eNOS genotype and plasma NO concentration, which was consistent with other studies (47, 48).
A limitation of the current study was that it was conducted in healthy, middle-aged individuals at moderate risk of CVD, who generally have low baseline concentrations of CVD risk markers and, thus, a lesser chance for improvement after LC n−3 PUFA supplementation (9). The dose and duration of the current trial could also have presented a limitation. However, 2 recent meta-analyses that evaluated effects of fish-oil supplementation on vascular and endothelial functions (9, 49), including 16 trials with doses that ranged from 0.45 to 4.5 g n−3 PUFAs/d and durations that ranged from 2 to 52 wk, concluded that there was no influence of either the dose or duration. There was a significant impact of study quality, whereby some poorer-quality studies appeared to be driving the beneficial effects of fish oil. However, it is notable that some of the highest-quality studies showed positive effects of n−3 PUFAs at doses as low as 0.84 g/d and a duration of only 2 wk (9, 49). The current study had a crossover design, which carried an inherent risk of a carry-over effect, although there was no indication that this was the case.
In conclusion, emerging cellular markers of endothelial damage, integrity, and repair appear to be sensitive to modification by dietary fatty acids. A moderate dose of LC n−3 PUFAs significantly increased numbers of circulating EPCs and decreased numbers of EMPs in the absence of alterations in blood vascular function, blood lipids, or circulating endothelial and inflammatory markers and independent of eNOS genotype.
Acknowledgments
We thank all volunteers who took part in the study and Kim Jackson for assistance with genotyping.
The authors’ responsibilities were as follows—PY, JAL, and ST: designed the study; SW and JM-P: conducted the research; PY: supervised the performance of the research; S-YW and JM-P: analyzed data and performed statistical analyses; JM-P and PY: wrote the manuscript; and all authors: read and approved the final manuscript. None of the authors declared a conflict of interest.
Footnotes
-
↵5 Abbreviations used: AI, augmentation index; AI@75, augmentation index adjusted for a heart rate of 75 beats/min; CVD, cardiovascular disease; EMP, endothelial microparticle; EPC, endothelial progenitor cell; FAME, fatty acid methyl ester; FMD, flow-mediate dilatation; KDR, kinase insert domain receptor; LC, long chain; mAb, monoclonal antibody; NO, nitric oxide; PMP, platelet microparticle; PPAR-γ, peroxisome proliferator–activated receptor γ PWA, pulse wave amplitude; RHI, reactive hyperemia index.
- Received March 26, 2014.
- Accepted August 5, 2014.
