Plasma and tissue long-chain (LC) PUFA concentrations are associated with the risk of several diet-related chronic diseases, including CVD⁽ Reference Bucher, Hengstler and Schindler ^{1 –} Reference Dacks, Shineman and Fillit ^{5 )}. Therefore it is important that the determinants of LC-PUFA metabolism, and concentrations in the circulation and in target tissues are fully understood. n-3 Fatty acids are PUFA, which contain the first double bond at the third carbon atom from the methyl end of the fatty acid. There are three major LC n-3 PUFA in the human diet and mammalian tissues, namely α-linolenic acid (αLNA), EPA and DHA. Although the most effective means to increase EPA and DHA status is through increased consumption of fish, bioconversion from the essential fatty acid, αLNA, represents a significant source and in particular in non-fish/EPA plus DHA supplement consumers who have 57–80 % lower intakes than fish eaters, with EPA and DHA derived from the sequential desaturation and elongation from αLNA⁽ Reference Welch, Shakya-Shrestha and Lentjes ^{6 )}.

The potential health benefits associated with consumption of EPA and DHA are numerous, with the most studied and accepted being a reduction in CVD risk. As summarised in several systematic reviews and meta-analysis of prospective epidemiological studies and randomised controlled trials, the ability of LC n-3 PUFA to reduce all-cause mortality and cardiovascular mortality has been widely described⁽ Reference Bucher, Hengstler and Schindler ^{1 ,} Reference Mente, de Koning and Shannon ^{2 ,} Reference Wang, Harris and Chung ^{4 ,} Reference León, Shibata and Sivakumaran ^{7 ,} Reference Studer, Briel and Leimenstoll ^{8 )}. However, it should be noted that this is not a fully consistent finding, with the heterogeneity in responsiveness as yet not fully understood⁽ Reference Rizos, Ntzani and Bika ^{9 ,} Reference Kwak, Myung and Lee ^{10 )}. Consumption of EPA and DHA has also been shown to be associated with many other diseases, for example, autoimmune diseases such as rheumatoid arthritis, cancer, diabetes, respiratory diseases, gastrointestinal diseases, Alzheimer's disease, depression, as well as psychotic disorders, for example schizophrenia⁽ Reference Fetterman and Zdanowicz ^{11 –} Reference Jump, Depner and Tripathy ^{14 )}.

The current recommended intakes for EPA plus DHA in the UK are ⩾450 mg/d^{( 15 )}. This recommendation is based largely on the cardiovascular benefits of these fatty acids and can be achieved by consuming two portions of fish per week, one of which should be oily^{( 15 )}. However, the estimated EPA and DHA consumption in adults in the UK is approximately 270 mg/d for men and 220 mg/d for women, which is far below the recommended minimal intake⁽ Reference Welch, Shakya-Shrestha and Lentjes ^{6 )}. Furthermore, mean population intakes are known to be highly skewed, with a large proportion of the population who do not consume fish or an EPA/DHA-containing supplement having a typical EPA plus DHA intake of <50 mg/d⁽ Reference Welch, Shakya-Shrestha and Lentjes ^{6 ,} Reference Minihane ^{16 )}.

n-6 PUFA, including linoleic acid (LA) and arachidonic acid (AA), contain the first double bond at the sixth carbon atom from the methyl end of the fatty acid. LA is an essential fatty acid that is found in vegetable oils and is the most abundant PUFA in the modern Western diet⁽ Reference Chilton, Murphy and Wilson ^{17 )}. LA can be metabolised to AA, which in turn, is a precursor of eicosanoids, such as PG, thromboxanes and leukotrienes. These eicoisanoids tend to be pro-inflammatory and therefore may negatively impact on the development of CVD⁽ Reference Martinelli, Girelli and Malerba ^{18 )}.

There is now a large published literature reporting on the impact of individual gene variants on LC-PUFA metabolism and CVD incidence and biomarker profiles. This review will focus on the fatty acid desaturase (FADS) genotypes, which are emerging as the most significant common genetic determinants identified to date. Accumulating evidence suggests that the locus may, in the future, be useful in stratification and targeting of LC-PUFA recommendations towards individuals likely to be deficient and responsive.

PUFA bioconversion and the fatty acid desaturase genotype

In addition to dietary intake, tissue EPA and DHA is influenced by the rate of bioconversion from αLNA, which involves multiple desaturation and elongation steps (Fig. 1). The Δ-5 and Δ-6 desaturase enzymes are the key rate-limiting enzymes in this pathway⁽ Reference Sprecher ^{19 )}. The human desaturase complementary DNA were first cloned in 1999 by Cho et al.⁽ Reference Cho, Nakamura and Clarke ^{20 ,} Reference Cho, Nakamura and Clarke ^{21 )} and were later identified as FADS1 and FADS2 in the human genome⁽ Reference Marquardt, Stohr and White ^{22 )}, located in a cluster on chromosome 11 (11q12–13.1). Δ-5 desaturase and Δ-6 desaturase are found in many human tissues, but the liver is the site at which they are most highly expressed⁽ Reference Cho, Nakamura and Clarke ^{20 ,} Reference Cho, Nakamura and Clarke ^{21 )}. LA and αLNA are metabolised by the same series of enzymes. EPA and DHA are produced at limited conversion rates of 0·2–6 % for EPA and <0·1 % for DHA in human males and post-menopausal females, with higher rates evident in pre-menopausal females⁽ Reference Burdge ^{23 )}. The more efficient EPA and DHA synthesis in pre-menopausal women is thought to be an evolutionary adaptation, so that younger females have sufficient LC-PUFA to meet the demands of pregnancy and the developing fetus. As will be described, variation across the FADS gene region appears to be important in modulating LC-PUFA status. The functional SNP in FADS1 and 2 have not yet been identified.

Fig. 1. Synthesis of long-chain PUFA from linoleic acid (LA) and alpha-linolenic acid (αLNA). Both LA (n-6) and αLNA (n-3) are elongated, desaturated and β-oxidised using the same enzyme system. AA, arachidonic acid.

Impact of fatty acid desaturase genotype on PUFA status

Using both a candidate gene (Table 1) and a genome wide association study (Table 2) approach, numerous studies have reported associations between variations in the FADS locus and desaturase activity and fatty acid status in human subjects. Desaturase activity can be approximated by calculating the product-to-precursor ratio of fatty acids. In 2006, Schaeffer et al.⁽ Reference Schaeffer, Gohlke and Muller ^{24 )} analysed eighteen SNP and reconstructed haplotypes in the FADS1–2 cluster in 727 adults. A five-locus FADS haplotype accounted for 27·7, 5·2 and 1·4 % of the variation in AA, EPA and DHA in serum phospholipids, respectively. The minor alleles were associated with higher αLNA and LA and lower γ-linolenic acid, AA, EPA and n-3 docosapentaenoic acid concentrations, with no significant impact on DHA⁽ Reference Schaeffer, Gohlke and Muller ^{24 )}. More recently, Ameur et al. performed genome wide genotyping in 5652 individuals, and targeted resequencing (n 960) of the FADS region, across five European population cohorts and reported that present-day human subjects have two common FADS haplotypes, which are defined by twenty-eight closely linked SNP, one of which was considered to be more efficient in relation to the biosynthesis of LC-PUFA⁽ Reference Ameur, Enroth and Johansson ^{25 )}. This FADS haplotype was associated with lower levels of LA (borderline significant) and αLNA and higher levels of EPA, γ-linolenic acid, DHA and AA. Over the last decade, a number of other candidate gene approach studies, as well as genome wide association studies, have been conducted and the association between FADS SNP/haplotypes and PUFA status, as well as desaturase activities, in plasma have been confirmed and extended to tissue fatty acid composition (Tables 1 and 2). However, information on how factors, including n-3 PUFA intakes, health status and ethnicity, may influence the penetrance of the FADS genotype, and in turn the effect size, is relatively unknown. Further research, expanding on the recent research by Wang et al.⁽ Reference Wang, Athinarayanan and Jiang ^{26 )}, is also required to determine the functional SNP, as well the molecular mechanism(s) responsible for the effect of the FADS genotype on EPA and DHA status. Wang et al. examined the association between six FADS SNP and the lipidomic profile and FADS1–3 expression in liver samples (n 154) and reported all six alleles to be associated with FADS1 (but not FADS2 and 3) gene expression and protein levels, suggesting that the causal variant(s) may be located at FADS1 ⁽ Reference Wang, Athinarayanan and Jiang ^{26 )}. In addition, twenty out of forty-two highly linked SNP were located in the transcription factor-binding sites of the locus. Although it is unclear exactly which SNP is causal and exactly how the SNP influences transcription factor binding and activation of FADS1, the findings add considerable credibility to the observations that FADS genotypes influence EPA and DHA status.

Table 1. Candidate gene studies: associations between fatty acid desaturase SNP and fatty acid status and cardiovascular health

AA, arachidonic acid; LA, linoleic acid; αLNA, alpha-linolenic acid, MI, myocardial infarction, CAD, coronary artery disease; hs-CRP, high sensitivity C-reactive protein; EDA, eicosadienoic acid; LD, linkage disequilibrium; D6D, Δ-6-desaturase; D5D, Δ-5-desaturase; DGLA, dihomo-γ-linolenic acid; HR, hazard ratio; ACS, acute coronary syndrome; VLCFA, very long-chained fatty acids; T2D, type 2 diabetes.

Table 2. Genome wide association studies: associations between fatty acid desaturase SNP and fatty acid status and cardiovascular health

PC, phosphatidylcholine; D5D, Δ-5-desaturase; AA, arachidonic acid; TC, total cholesterol; LA, linoleic acid; αLNA, alpha-linolenic acid; GLA, γ-linolenic acid; DGLA, dihomo-γ-linolenic acid; TFA, trans-fatty acid.

Impact of diet composition on the relationship between the fatty acid desaturase genotype and PUFA and cardiovascular health status

There have been a number of studies that show that diet composition can influence the relationship between FADS genotype and plasma fatty acid and lipid status (Table 3). In 2012, Hellstrand et al. reported that the FADS rs174547 minor allele was associated with lower LDL-cholesterol among individuals in the lowest tertile of LC n-3 PUFA intakes⁽ Reference Hellstrand, Sonestedt and Ericson ^{38 )}. A significant interaction between rs174547 and the ratio of αLNA and LA intakes on HDL-C was also observed⁽ Reference Hellstrand, Sonestedt and Ericson ^{38 )}. More recently, a 14-year follow-up in 24 032 participants reported that the αLNA:LA intake ratio was inversely associated with CVD risk only among participants homozygous for the rs174547 minor allele⁽ Reference Hellstrand, Ericson and Gullberg ^{39 )}. αLNA intakes were also inversely associated with ischaemic stroke in this genotype group. In addition to observational analysis, the impact of FADS variants on response to LC-PUFA supplementation has also been examined. Gillingham et al. carried out a randomised crossover trial in thirty-six hyperlipidemic subjects in which three diets (enriched with flaxseed oil or high-oleic acid canola oil compared with a typical Western diet) were consumed for 4 weeks and five FADS SNP were analysed⁽ Reference Gillingham, Harding and Rideout ^{40 )}. Subjects with minor allele variants (rs174545, rs174583, rs174561, rs174537) had decreased desaturase activity, but an increase in αLNA intakes resulted in greater increases in plasma EPA than in major allele homozygotes consuming αLNA intakes typical of a Western diet⁽ Reference Gillingham, Harding and Rideout ^{40 )}. Cormier et al. conducted a study in 208 subjects examining the impact of fish-oil supplementation (1·9–2·2 g/d EPA and 1·1 g/d DHA) for 6 weeks and nineteen FADS SNP on plasma TAG and reported that rs174546 was associated with TAG, but no significant genotype by supplementation interaction was observed⁽ Reference Cormier, Rudkowska and Paradis ^{41 )}. In terms of whole-diet interventions, one study to date has examined the interaction of FADS genotype and the Mediterranean diet on serum and colonic fatty acid profiles⁽ Reference Porenta, Ko and Gruber ^{42 )}. In a 6-month intervention (n 108) and genotyping for four FADS SNP, a significant diet by genotype interaction for AA concentrations in the colon was observed; subjects with FADS major alleles following the Mediterranean diet had 18 % lower AA concentrations than subjects on the control diet (healthy eating diet)⁽ Reference Porenta, Ko and Gruber ^{42 )}. There were no significant diets by genotype interactions for other colonic or serum fatty acids. Overall, it is clear that further research is necessary to determine the potential of the diet, particularly dietary fatty acids, to modify the relationship between the FADS genotype and fatty acid status. An investigation of diet composition × FADS genotype × fatty acid status represents a secondary objective of the recently completed NU-AGE intervention.

Table 3. Diet–gene interactions: impact of dietary intakes/interventions on associations between fatty acid desaturase genotype and fatty acid status and cardiovascular health

TC, total cholesterol; LC, long chained; αLNA, alpha-linolenic acid; LA, linoleic acid; DPA, docosapentaenoic acid; hs-CRP, high sensitivity C-reactive protein; HR, hazard ratio; D5D, Δ-5-desaturase; D6D, Δ-6-desaturase; LDL-C, LDL-cholesterol.