Ethnicity, sex, FADS genetic variation, and hormonal contraceptive use influence delta-5- and delta-6-desaturase indices and plasma docosahexaenoic acid concentration in young Canadian adults: a cross-sectional study

Background There is great interest in the relationship between polyunsaturated fatty acids and health. Yet, the combinatory effect of factors such as sex, ethnicity, genetic polymorphisms and hormonal contraceptives (HC) on the concentrations of these fatty acids is unknown. Therefore, we sought to determine the effects of FADS polymorphisms, and HC use in females, on aggregate desaturase indices (ADI), and plasma docosahexaenoic acid (DHA) concentrations in Caucasian and East Asian males and females. Methods Fasting plasma samples were collected from subjects (Caucasian males: 113 and females: 298; East Asian males: 98 and females: 277) from the Toronto Nutrigenomics and Health Study. Fatty acid concentrations were measured by gas chromatography. ADI were estimated by dividing concentrations of arachidonic acid by linoleic acid (n-6 ADI) and eicosapentaenoic acid (EPA) by α-linolenic acid (n-3 ADI). [DHA/EPA] desaturase index was used to determine effects of FADS2 polymorphisms and HC use on EPA conversion to DHA. Results In Caucasians, associations between n-6 ADI and multiple SNP (FADS1 rs174547, FADS2 rs174576, and rs174611 in males; FADS1 rs174547, FADS2 rs174570, rs174576, rs174679, rs174611, rs174593, rs174626, rs2072114, rs2845573, and rs2851682 in females) withstood multiple testing. In East Asian females, 5 SNP-n-6 ADI associations (FADS2 rs174602, rs174626, rs2072114, rs2845573, and rs2851682) withstood multiple testing. One FADS2 SNP was associated with altered [DHA/EPA] desaturase index in Caucasian females only (rs174576, p < 0.0001). HC use had a significant effect on DHA concentrations in Caucasian females only (P < 0.0001). Conclusions We demonstrate ethnic- and sex-specific effects of FADS polymorphisms on desaturase indices, and ethnic-specific effect of HC use on plasma DHA concentrations. Electronic supplementary material The online version of this article (doi:10.1186/s12986-015-0010-9) contains supplementary material, which is available to authorized users.

In recent years, a large body of work has addressed the effects of n-3 and n-6 PUFA on different aspects of our health (reviewed in [6]). Many of the reported biological effects of n-3 and n-6 PUFA seem to be mediated primarily by AA, EPA and DHA [6]. AA is a precursor of both anti-inflammatory, pro-inflammatory, pro-atherogenic and prothrombotic mediators, while EPA and DHA have been shown to play a role in cognitive development and in protection against cardiovascular disease and inflammatory conditions [7][8][9][10].
Several factors have been implicated in the alteration of endogenous fatty acid levels. These include diet, genetic variations and hormonal regulation [11,12]. Single nucleotide polymorphisms (SNP) in FADS1 and FADS2 have been shown to alter circulating FA levels reflected in changes in corresponding estimates of D5D and D6D enzymatic activity [13][14][15]. Moreover, FADS genetic variations have been shown to be associated with increased HOMA-IR and insulin resistance as well as hsCRP levels; thus, FADS polymorphisms can play a role in modifying disease risk [16,17]. In addition, another modifier is hormonal contraceptives (HC) which have been positively correlated with DHA levels in women [18,19].
We have previously shown the differential influence of FADS polymorphisms on D5D and D6D desaturation indices depending on ethnicity (Caucasian vs. East Asian) [20]; however, differential effects of FADS genetic variations based on sex are yet to be determined. In addition, the effects of hormonal contraceptive use on endogenous DHA concentrations in different ethnicities are also unknown. Identification of factors that differentially influence circulating fatty acid concentrations will advance our understanding of sex-and ethnic-specific aspects of fatty acid metabolism. Thus, in this study we sought to determine the differential effects of FADS polymorphisms and HC use on desaturase indices and circulating DHA concentrations in Caucasian and East Asian males and females.

Study population
Participants (total n = 786; Caucasians: males = 113, females = 298; East Asians: males = 98, females = 277) between the ages of 20-29 were recruited as part of the cross-sectional Toronto Nutrigenomics and Health (TNH) Study [21] between September 2004 and July 2009. Written informed consent was obtained from all participants. Subjects were free-living and were required to fast overnight for a minimum of 12 h prior to collection of blood. Standard clinical procedure was followed for the measurement of glucose, insulin, total-, and HDL-cholesterol, triglycerides, and free fatty acids [21]. HOMA-IR was calculated using the homeostasis model assessment method [22]. For the purpose of fatty acid analysis, plasma was separated from blood samples by centrifugation. Plasma samples were then frozen and stored at −80°C. Anthropometric measurements were recorded for all participants and health, life style, and food frequency questionnaires were completed by subjects. Completed questionnaires were used to calculate total energy intake from fat and physical activity scores, as previously described [21,23]. Ethnicity (Caucasians and East Asians) was the only inclusion criterion. Subjects were not excluded if they were fish or supplement consumers; however, to address dietary omega-3 and omega-6 intake regression models were adjusted for intake levels of LA, ALA, EPA and DHA obtained from FFQ. Women who were pregnant or breastfeeding were not included in the study. No other exclusion criteria were included in these analyses. The study protocol was approved by the Research Ethics Boards at the University of Toronto and University of Guelph.

Gas chromatography analysis
Plasma total lipids were extracted and analyzed as described previously [20,24]. Free fatty acid C17:0 was used as an internal standard (5 μg of 1 mg/ml stock) and was used to calculate fatty acid concentrations (μg/ml). N-3 and n-6 aggregate desaturase indices (ADI) were calculated by dividing concentrations of EPA by ALA and AA by LA, respectively [20]. [DHA/ALA] and [DHA/EPA] desaturase indices were calculated by dividing concentrations of DHA by ALA and DHA by EPA, respectively. These indices were used to investigate the effects of FADS1/2 polymorphisms and HC use on conversion of ALA to DHA and EPA to DHA, respectively.

Genotyping
Identification of FADS1/FADS2 SNP, selection of SNP and genotyping was performed as described previously [20]. Briefly, SNP in FADS1 and FADS2 were identified independently with the International HapMap Project SNP database. Tag SNP (tSNP) were selected with a minor allele frequency (MAF) greater than 0.05 and pairwise tagging (r 2 ≥ 0.8), leading to the identification of 19 tSNP in total (3 in FADS1 and 16 in FADS2 that captured more than 43 initially identified SNP). Linkage disequilibrium (LD) was examined between the 19 tSNP identified using the SNP Annotation and Proxy Search (SNAP) database. Two tSNP were in high LD (r 2 ≥ 0.8): rs174547 in FADS1 and rs174576 in FADS2. Sequenom MassARRAY platform was used for genotyping, which is based on detection through MALDI-TOF MS (Mass Array, Sequenom, San Diego, CA). All 19 tSNP were replicated in 29 randomly selected DNA samples and 100% concordance was achieved [17].

Statistical analysis of data
Results are expressed as mean ± standard error mean (SEM). Data analysis and testing for Hardy-Weinberg equilibrium was carried out using JMP genomics software V5 (SAS Institute, Cary, NC). A student's t-test was used to determine differences in fatty acid concentrations and ratios between males and females. A Tukey's Honestly Significant Difference post-hoc test was used to determine differences in desaturase indices for each genotype. P-values of analyses of fatty acid levels or desaturase indices were determined using multiple linear regression models which were adjusted for BMI, age, dietary LA, ALA, EPA and DHA, % total energy from dietary fat and physical activity. Multiple linear regression models were used to identify associations between individual desaturase indices and SNP or HC use and were adjusted for BMI, age, dietary LA, ALA, EPA and DHA, % total energy from dietary fat and physical activity. Due to the high LD known to exist between SNP in the FADS gene cluster, the SNP with the strongest association (i.e., smallest p-value) was subsequently included as a covariate in repeat linear regression models of identified SNP-ADI associations to determine whether these associations were dependent or independent of the covariate SNP. If the originally identified association was lost after addition of the covariate SNP to the model, the effect of the SNP investigated was considered dependent on covariate SNP. If the association remained significant, the effect of the SNP investigated was considered to be independent of the covariate SNP. A p-value of < 0.05 was considered statistically significant. To determine p-value cut off for multiple testing, 0.05 was divided by the number of total linear regression analyses performed for associations of SNP and desaturase indices (160 analyses) which resulted in a significant p-value cut off of 0.0003.

Study population
Within Caucasian males and females, there were no significant differences in BMI, energy intake from fats or levels of triglycerides and free fatty acids. However, females had significantly higher values of HOMA-IR and higher levels of insulin, total cholesterol, HDL-cholesterol and LDL-cholesterol and significantly lower levels of glucose compared to males (Table 1). In East Asian males and females, there were no significant differences in values of HOMA-IR and energy intake from fats or levels of insulin, LDL-cholesterol, triglycerides and free fatty acids. There were significantly higher levels of total cholesterol and HDL-cholesterol in females compared to males and lower BMI and glucose levels in females (Table 1).

PUFA concentrations and desaturation indices N-6 PUFA
Concentrations of selected n-6 and n-3 PUFA are presented in Table 2

Sex-specific effects of FADS SNP on desaturase indices up to AA and EPA Caucasian males
Multiple linear regression analysis showed 2 associations between n-3 ADI and FADS genetic variants (rs174611 and rs526126) in Caucasian males; however, none withstood multiple testing.
Multiple linear regression analyses of the 19 SNP and n-6 ADI revealed significant associations with 13 SNP (rs174547, rs174570, rs174576, rs174602, rs174593, rs174611, rs174626, rs174627, rs17831757, rs2072114, rs2851682, rs526126, and rs968567) ( Table 3); however, only 3 SNP (rs174547, rs174576, and rs174611) withstood multiple testing. Carriers of the minor alleles of all aforementioned SNP were observed to have lower ADI. Additional analyses including the SNP with the strongest association (rs174547, p = 1.57 × 10 −6 ) as a covariate resulted in the loss of the association between n-6 ADI and rs174611. In addition, results suggested that effects of rs174547 and rs174576 were dependent on each other.

Caucasian females
Multiple linear regression analyses of the 19 SNP and n-3 ADI in Caucasian females revealed significant associations only between n-3 ADI and 7 SNP (rs174547, rs174576, rs174579, rs174626, rs174627, rs174593, and rs968567) (Table 4); however, adjustment for multiple testing resulted in the loss of all associations.

East Asian males
No significant associations between SNP and n-3 or n-6 ADI were identified in East Asian males (Table 5).

DHA synthesis from EPA and influence of FADS2 SNP
Associations between FADS polymorphisms and [DHA/ ALA] desaturase index were determined but no significant associations withstood multiple testing (data not shown).

Caucasian population
Multiple linear regression analyses of [DHA/EPA] desaturase index and FADS2 SNP were performed. Only rs526126 was significantly associated with altered [DHA/EPA] in Caucasian males; however, the association did not withstand multiple testing. In Caucasian females rs174570, rs174576, rs174593, rs2072114 and rs2845573 were significantly associated with altered [DHA/EPA] desaturase index (Table 7). rs174576 displayed the strongest association with the [DHA/EPA] desaturase index (p < 0.0001) and was the only SNP that remained significant after adjustment for multiple testing. The minor allele, A, was associated with increased [DHA/ EPA] desaturase index in Caucasian females.

Asian population
In East Asian males, 3 SNP (rs2072114, rs2851682 and rs526126) were significantly associated with altered [DHA/EPA] desaturase index (Table 8); however, none of the associations withstood multiple testing.

Discussion
In this study, we undertook a comprehensive examination of major factors that influence the metabolism of   essential fatty acids from precursor to product. Sex, ethnicity and FADS polymorphisms have been shown to play important roles influencing the conversion of ALA to EPA and LA to AA. However, the present study revealed that genetic polymorphisms have less influence in the metabolism of EPA to DHA while HC use plays a prominent role in that process, specifically in Caucasian women. These findings highlight the differential roles of sex, ethnicity, and FADS polymorphisms within discrete steps in the metabolic conversion of essential fatty acids.

Role of FADS polymorphisms on [EPA/ALA] and [AA/LA] desaturation indices
Previously, we showed that FADS1 and FADS2 genetic variants alter desaturase indices in Caucasians and East Asian Canadian adults [20]. These analyses were examined in a smaller mixed gender subset of the TNH subjects (147 subjects) while the present study examined sex-specific associations in 786 subjects that met our inclusion and exclusion criteria. Findings from the present study confirmed significant associations previously reported by our group showing that n-6 ADI are associated with FADS1 and FADS2 polymorphisms (rs174547, rs174576) in Caucasians. Using the SNP Annotation and Proxy Search (SNAP) database, rs174547 and rs174576 were found to be in high linkage disequilibrium [20]. We also confirmed that rs2072114 is strongly associated with n-6 ADI in East Asians; however, examination of sex-dependent associations in this study, revealed that the rs2072114-ADI association is only significant in females. An additional observation is that the minor alleles of the identified FADS SNP (rs174547, rs174570, rs174576, rs174679, rs174593, rs174602, rs174611, rs17 4626, rs2072114, rs2845573, rs2851682) were associated with lower [AA/LA], but not [EPA/ALA], desaturation indices. The potential significance for carriers of the minor alleles are lower [AA/LA] and higher [EPA/ALA] indices, subsequently, lower risk of inflammatory diseases associated with the production of lower amounts of AA and higher amounts of EPA derived inflammatory and anti-inflammatory eicosanoids, respectively [17]. FADS polymorphisms have been shown to explain 28% of variability in AA levels in serum phospholipid fatty acids of Caucasian participants [25]. AA is a primary precursor for the synthesis of pro-inflammatory eicosanoids as well as the vasoconstrictor TXB 2 , and AA-   generated prostaglandins are shown to be pro-arrhythmic [26][27][28]. Taken together, the identified association between SNP and altered n-6 ADI in females may contribute to modified risk of inflammatory and cardiovascular conditions through the alteration of circulating concentrations of AA. We acknowledge that a limitation of our study population is that it is comprised of young healthy Canadian adults but differences of circulating fatty acids may be important early biomarkers in chronic disease prevention.

Role of FADS2 polymorphisms in altered [DHA/EPA] desaturation index and DHA concentrations between sexes and ethnicities
Previously, the impact of FADS polymorphisms on D5D and D6D desaturase indices was investigated in regard to AA and EPA synthesis [20]. However, D6D is also involved in the synthesis of DHA from EPA [5]. Therefore, we sought to determine the role of FADS2 genetic variations on D6D desaturase index in regard to subsequent conversion of EPA to DHA. Others have reported significant associations between FADS1/FADS2 polymorphisms with n-6 and n-3 fatty acids except for DHA [15,25]. The present study identifies a significant association of one  [30]), were identified as the driving factor in the more efficient conversion of ALA to DHA in females [12]. Previously, increased DHA synthesis was observed in females using HC and, more recently, HC use was positively correlated with percent DHA of total fatty acids in red blood cell of Icelandic women [18,19]. In this regard, our study showed significantly higher [DHA/EPA] desaturase indices in females, and in particular, HC users, relative to males (Figure 2 [31]. It can also be argued that the lower concentrations of DHA in Caucasians, as a result of lower dietary DHA intake, allowed for the significant effect of both FADS2 SNP and hormonal contraceptives to be unmasked. It is noteworthy that further stratification of HC users by type of HC used was not performed. A study examining effects of HC use on plasma proteomic biomarkers in TNH study subjects found that only the use of HC significantly altered the proteomic profile while the type of HC or duration of use had no significant effect [32]. The study also found that some of the proteins affected by HC use were biomarkers of dysregulation of inflammation, which is consistent with other reports [32][33][34]. Given that the findings of the present study can only show a relationship exists and not causality, (+) and (−) signs are used to denote relative differences in conversion (high and low, respectively) indirectly measured as a ratio of product to precursor fatty acid. Abbreviations: HC hormonal contraceptives, SNP single nucleotide polymorphisms, n/a not applicable, n/s no significant effect.
DHA has been shown to be protective against a range of chronic diseases; therefore, it is important to understand factors that may render some women more susceptible to low levels of this fatty acid. Furthermore, DHA has a known role in cognitive development; thus, it is necessary to identify factors (such as ethnicity, FADS polymorphisms, or history of HC use) that may influence mothers circulating levels of DHA that is passed to their offspring.

Strengths, weaknesses, and overall conclusions
We acknowledge the following limitations in this study: First, the use of plasma samples for measurement of endogenous fatty acids concentrations. Red blood cells and adipose tissue reflect longer term dietary fatty acid intake; however, fasting plasma fatty acids as an accessible tissue correlate with adipose tissue fatty acids [35]. Plasma fatty acids levels are determined by hepatic synthesis and diet; however, since subjects fasted overnight prior to sample collection, contribution of diet to plasma fatty acids concentrations is expected to be insignificant. Therefore, confounding by recent dietary fat intake is minimized. We also acknowledge that phospholipids are more frequently used for the investigation of polyunsaturated fatty acids (PUFA); however, studies that measure fatty acids from phospholipids often report and use percentage composition. In this study we have determined fatty acids concentrations (as opposed to percentage composition values), therefore, provided a quantitative perspective of specific circulating fatty acids, rather than relative changes that are dependent on levels of other fatty acids. The use of the desaturase index as an indirect measure of fatty acid conversion has been widely utilized [13,[36][37][38]. While endogenous conversion by desaturases is recognized to be low, the desaturase index appears to be sufficiently sensitive across a number of studies [13,[36][37][38].

Conclusions
This study identified sex-and ethnic-specific effects of FADS polymorphisms on desaturase indices (Table 10); thus, demonstrating the importance of stratification of sample populations investigated by sex, ethnicity, genetics and contraceptive use. The increasing recognition that n-6 and n-3 PUFA influence chronic disease risk validate the need for better understanding of factors influencing circulating levels of these fatty acids.

Additional file
Additional file 1: Table S1. HWE values of FADS1 and FADS2 and their genotype frequencies. Abbreviations