High triacylglycerol (TG) levels are associated with cardiovascular disease (CVD) [1]. Population mean TG levels have increased since 1976 in parallel with the constant growing epidemic of obesity, insulin resistance and type-2 diabetes mellitus [2,3]. A meta-analysis of 17 population-based prospective trials including 46,413 men and 10,864 women identified plasma TG levels as an independent risk factor of CVD and estimated that for each increase of 1.0 mmol/L, the relative risk of CVD increased approximately by 30% for men and 75% for women [4,5].

Numerous studies have demonstrated the beneficial effects of omega-3 (n-3) polyunsaturated fatty acids (PUFA), especially eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) on reducing CVD risk factors [6,7,8,9]. The intake of EPA and DHA has been associated with a reduced risk of myocardial infarction and the prevalence of recurrence [10,11]. A review of human studies reported that 4 g of marine-derived n-3 PUFA per day decreased plasma TG concentrations by 25% to 30% [4]. The American Heart Association recommends an intake of 2 to 4 g of EPA/DHA per day for patients who need to lower their plasma TG levels [12]. The decreased in plasma TG that is observed with high intakes of n-3 PUFA appears to be secondary to the increased hepatic β-oxidation and decreased lipogenesis [13].

The Fish Oil Intervention and Genotype (FINGEN) study showed that TG levels of 31% of the volunteers did not show any reduction after an 8-week supplementation with 1.8 g/day of EPA + DHA [14]. Innate characteristics such as gender, age and genetic factors may contribute to the variability in benefits reported from intervention trials using n-3 PUFAs [15]. Therefore, the anticipated effect of such a supplementation on an individual basis does not necessarily match those reported for the general population [16].

The large inter-individual variability observed in the plasma lipid response to a supplementation with n-3 PUFA may partly result from genetic variations. Recent data suggest that single-nucleotide polymorphisms (SNPs) found in genes involved in metabolic pathways of n-3 PUFA contribute to the variability of PUFA levels [15,17]. The fatty acid desaturase-1 (FADS1) and fatty acid desaturase-2 (FADS2) genes encode respectively for two desaturases: δ-5 desaturase (D5D) and δ-6 desaturase (D6D) [17]. The D5D and D6D, responsible for double bonds formation in the n-3 PUFA pathways, have been associated with differences in fatty acid (FA) composition of plasma [18], erythrocyte membranes [19] and adipose tissue [18]. Another potential desaturase, whose function remains to be elucidated, is possibly encoded by the fatty acid desaturase-3 gene (FADS3) [20].

Concerns regarding the efficiency of n-3 PUFA supplementation remain. The TG lowering effects of n-3 PUFA using several ratios and doses of EPA and DHA have been reported in different studies. However, these studies do not allow setting the optimal conditions. Conflicting data exist and may arise from inter-individual genetic variations. The purpose of the present study is to test whether the plasma lipid response to a 6-week n-3 PUFA supplementation is influenced by genetic variations in the FADS gene cluster.

Body weight, height, and waist girth were measured according to the procedures recommended by the Airlie Conference [22] and were taken before the run-in period, as well as pre- and post-n-3 PUFA supplementation. BMI was calculated as weight in kilograms divided by height in meters squared (kg/m²).

Blood samples were collected from an antecubital vein into vacutainer tubes containing EDTA after 12 h overnight fast and 48 h alcohol abstinence. Blood samples were taken to identify and exclude individuals with any metabolic disorders. Afterwards, participants had blood samples taken prior to and after the n-3 PUFA supplementation period. Plasma was separated by centrifugation (2500× g for 10 min at 4 °C) and samples were aliquoted and frozen for subsequent analyses. Plasma total cholesterol (TC) and TG concentrations were measured using enzymatic assays [23]. The high-density lipoprotein cholesterol (HDL-C) fraction was obtained after precipitation of very low-density lipoprotein and low-density lipoprotein particles in the infranatant with heparin manganese chloride [24]. Low-density lipoprotein cholesterol (LDL-C) was calculated with the Friedewald formula [25]. Apolipoprotein B-100 (ApoB100) concentrations were measured in plasma by the rocket immunoelectrophoretic method of Laurell, as previously described [26]. Plasma C-reactive protein (CRP) was measured by nephelometry (Prospec equipment Behring) using a sensitive assay, as described previously [27].

Data were analyzed with SAS statistical software V9.2 (SAS Institute, Cary, NC, USA) The ALLELE procedure was used to verify the departure from Hardy-Weinberg equilibrium (HWE) and to calculate minor allele frequency (MAF). Variables not normally distributed were log-transformed before analyses. ANOVA and the type III sum of squares were used to look for significant differences in daily energy and nutrient intakes, at prior and after an n-3 PUFA supplementation when age, sex and BMI were included in the model and to test for differences in plasma TG levels among groups divided on the basis of the genotype for rs174546. The repeated MIXED procedure was used to test for the effects of the genotype, the supplementation and the genotype by supplementation interaction on plasma TG and gene expression levels when age, sex and BMI were included in the model. Statistical significance was defined as p ≤ 0.05. To identify potential effects of variations located in the FADS gene cluster region, a transcription factor search was performed using MatInspector 8.0 software from the Genomatix Suite.

All SNPs were in HWE except two: rs7935946 and rs174579 (see Figure 1 for the LD plot). These SNPs were not considered for further analysis. Therefore, associations with 17 SNPs were tested in statistical analyses. The % gene coverage with these 17 SNPs was of 87%.

Baseline characteristics of study participants are presented in Table 2. According to these results, men and women were overweight (BMI > 25 kg/m²) and had mean plasma TG levels slightly above the cut-point value of 1.129 mmol/L recommended by the AHA for optimal plasma TG levels [29]. Gender differences are evident with respect to weight, TC/HDL-C ratio, CRP, HDL-C and TG levels.

Daily energy intakes were calculated by a food frequency questionnaire validated for healthy French-Canadian men and women [21] and are presented in Table 3. After the supplementation, carbohydrates, saturated fats and proteins were significantly different from the pre n-3 PUFA period (p = 0.0005, p = 0.0008 and p = 0.02, respectively). After the supplementation, the PUFA intake—taking into account food intakes and fish oil capsules used during the supplementation—was significantly increased (p = 0.003). In our study, the 6-week average of fish servings/week (a serving = 75 g of fish) was 0.89 servings/week based on the compliance questionnaire given at the end of the study. Furthermore, subjects were asked to limit their fish consumption to no more than 2 servings/week (maximum of 150 g). Based on these results, subjects who had consumed the maximum quantity of fish permitted per week would have had an extra 0.43 g of EPA + DHA/day. With the fish oil supplementation, the total of EPA + DHA was 3 g/day. Common food items eaten by the subjects were seafood, tuna, trout, haddock and salmon.

To test the potential interaction between the FADS gene polymorphisms and the n-3 PUFA supplementation on plasma TG levels, the MIXED procedure was used in order to test whether the genotype, the supplementation or the interaction (genotype by supplementation) were associated with plasma TG levels. As shown in Table 4, independently of the genotype, the supplementation was associated with fasting plasma TG concentrations meaning that the supplementation had an independent effect on plasma TG levels, as expected. One SNP was associated with plasma TG concentrations, rs174546, suggesting that this polymorphism modulates plasma TG levels. No significant genotype by supplementation interaction effects were observed. Further analyses revealed that, in the pre-supplementation period, plasma TG levels were lower in CC homozygotes when compared to carriers of the minor T allele (Table 5, p = 0.002). In the post-supplementation period, both genotype groups significantly decreased their plasma TG levels (Figure 2). However, there was no significant difference in post-supplementation plasma TG levels between the genotype groups when age, sex and BMI were included in the model (Table 5). Results remained unchanged after further adjustment for pre-supplementation plasma TG levels (data not shown). Since there was no significant difference between the two groups in post-supplementation plasma TG levels and since the interaction term was not significant in the repeated model, these results suggest that the FADS rs174546 is associated with plasma TG levels only and not with the plasma TG response to the n-3 PUFA supplementation.

In a model testing the effect of the genotype, the supplementation or the interaction (genotype by supplementation) on FADS1 gene expression, the SNP rs174546 was associated with FADS1 gene expression (p = 0.01) after adjustments for age, sex and BMI (data not shown). No effect of the n-3 PUFA supplementation (p = 0.54) and no gene by supplementation effect (p = 0.56) explained FADS1 gene expression.

In this study, we tested whether plasma TG levels during an n-3 PUFA supplementation varied according to the presence of common polymorphisms in the FADS gene cluster. 19 SNPs were initially chosen from the FADS gene cluster area covering all common variations of the FADS gene cluster. After genotyping, 17 SNPs were in HWE and thus were analysed in the present study. The FADS gene cluster area has been chosen due to the role of D5D and D6D activity in the n-3 PUFA metabolic pathway. D5D and D6D are essential parts of PUFA biosynthesis that catalyze a series of desaturation processes [30]. These desaturases are respectively encoded by FADS1 and FADS2. Also, a Genome-Wide Association Study has shown that the strongest evidence for an association of genetic contributors of plasma PUFA concentrations was observed in the FADS gene cluster area [31]. Some polymorphisms come from intergenic regions (rs174627, rs12807005. rs174448 and rs7482316) and are part of the FADS1 and FADS2 gene promoters because of the head to head orientation. FADS3 promoter was not considered since no desaturase activity is reported (see all selected SNPs in Table 1). All SNPs were polymorphic for the selected study population and were not in strong LD with each other.

In the present study, we observed an independent genotype effect of the SNP rs174546 on plasma TG levels and on FADS1 gene expression levels in a model including the SNP, the supplementation effect and the SNP by supplementation interaction. In the literature, SNP rs174546 has been much studied. Numerous studies have attributed beneficial effects to this polymorphism. Indeed. Dumont et al. showed that the minor allele of rs174546 was associated with decreased plasma TC and non-HDL-C levels [32]. In another study, Lu et al. reported similar results where the common C allele was associated with higher levels of TC non-HDL-C and HDL-C levels, but only in individuals consuming high intakes of omega-6 (>5.26% of total energy intake) [33]. Another study demonstrated that rs174547, in strong-LD with rs174546 (R² = 1.0), was a dominant SNP in the FADS gene cluster that influences desaturase activity and some evidence emerging from human-based research demonstrates that genetic variation in human desaturase genes affects enzyme activity and, consequently, disease risk factors [34]. They also showed that homozygotes for the major allele had a higher estimate of aggregate desaturase activity (ADA) reflected in n-3 PUFA while homozygotes for the minor allele had the lowest ADA [34]. This SNP is also associated with altered desaturase activity reflected in n-6 PUFA [34]. Recent studies have also shown that two common and very distinct FADS haplotypes were strongly associated with long-chain PUFAs synthesis levels. Haplotypes A and D, which includes rs174546, may exert differences in transcription levels and the ability to synthesize essential omega-3 and omega-6 long-chain PUFAs [35]. Reduced substrate (i.e., FA) availability leads to a reduction of VLDL TG synthesis [13]. FADS gene cluster, especially rs174546, is associated with TG levels and also to ADA, making it a significant SNP when talking about associations between lipids and FA metabolism.

Some SNPs may modulate desaturase activity and lead to changes in n-3 PUFA metabolism. We tested SNP rs174546 for potential functional significance using MatInspector 8.0 software, but it did not seem to alter transcription factor binding sites.

After the n-3 PUFA supplementation, considerable inter-individual variability in plasma TG levels was observed. It appears that some individuals require higher doses to achieve demonstrable benefits, whereas others are highly sensitive to relatively low doses and individuals with certain genotypes may experience adverse responses with respect to specific risk biomarkers, at least at high doses of n-3 PUFA [36]. FADS gene expression may modulate specific risk biomarkers in relation to certain genotype. Overall, the n-3 PUFA supplementation had no effect on FADS gene expression. The SNP rs174546 was a significant predictor of FADS1 gene expression levels. Caslake et al. showed that 31% of all volunteers had no reduction in plasma TG levels after 1.8 g/day of EPA + DHA for 8 weeks [14]. We basically observe similar results in the present study, as 28.8% of the participants had no reduction in plasma TG levels after the 6-week n-3 PUFA supplementation. These results demonstrate that intra-individual variability of plasma lipid levels is an important potential source of error enhancing the importance of genetic testing to identify individuals that are more likely to benefit from such therapies [37].

This study presents some limitations. Regarding daily intakes, significant differences for carbohydrates, saturated fats or proteins could be due to recall bias from subjects. Since subjects were asked to follow recommendations drawn from the Canada’s Food Guide to Healthy Eating, they could have reported food consumption differences that slightly changed calculated intakes. However, those differences did not affect significantly BMI nor energy intakes. Because carbohydrates intakes were significantly decreased in the post-supplementation period and this could impact on plasma TG levels, the carbohydrate intakes have been added into the statistical model and results remained unchanged (data not shown).

In summary, our data support the notion that the FADS gene cluster, especially SNPs from FADS1, are major determinants of plasma TG levels. SNP rs174546 may be an important SNP in the FADS gene cluster associated with plasma TG levels and FADS1 gene expression independently of a nutritional intervention with n-3 PUFA. These results need to be replicated in other independent studies. A better understanding of the phenomenon could allow the development of personalized dietary advice for prevention of CVD.