Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The role of a FADS1 polymorphism in the association of fatty acid blood levels, BMI and blood pressure in young children—Analyses based on path models

  • Maike Wolters ,

    Roles Conceptualization, Writing – original draft

    wolters@leibniz-bips.de

    Affiliation Leibniz Institute for Prevention Research and Epidemiology—BIPS, Bremen, Germany

  • Carmen Dering,

    Roles Formal analysis

    Affiliation Leibniz Institute for Prevention Research and Epidemiology—BIPS, Bremen, Germany

  • Alfonso Siani,

    Roles Investigation, Writing – review & editing

    Affiliation Epidemiology and Population Genetics, Institute of Food Sciences, National Research Council, Avellino, Italy

  • Paola Russo,

    Roles Investigation, Writing – review & editing

    Affiliation Epidemiology and Population Genetics, Institute of Food Sciences, National Research Council, Avellino, Italy

  • Jaakko Kaprio,

    Roles Validation, Writing – review & editing

    Affiliation Institute for Molecular Medicine FIMM, University of Helsinki, Helsinki, Finland

  • Patrizia Risé,

    Roles Investigation, Writing – review & editing

    Affiliation DiSFeB, Department of Pharmacological and Biomolecular Sciences, University of Milan, Milan, Italy

  • Luis A. Moreno,

    Roles Investigation, Writing – review & editing

    Affiliation GENUD (Growth, Exercise, Nutrition and Development) Research Group, Instituto Agroalimentario de Aragón (IA2), Instituto de Investigación Sanitaria Aragón (IIS Aragón), Centro de Investigación Biomédica en Red Fisiopatología de la Obesidad y Nutrición (CIBERObn), University of Zaragoza, Zaragoza, Spain

  • Stefaan De Henauw,

    Roles Investigation, Writing – review & editing

    Affiliation Department of Public Health, Faculty of Medicine and Health Sciences, Ghent University, Ghent, Belgium

  • Kirsten Mehlig,

    Roles Investigation, Writing – review & editing

    Affiliation Section for Epidemiology and Social Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

  • Toomas Veidebaum,

    Roles Investigation, Writing – review & editing

    Affiliation National Institute for Health Development, Tallinn, Estonia

  • Denés Molnár,

    Roles Investigation, Writing – review & editing

    Affiliation National Institute of Health Promotion, University of Pécs, Pécs, Hungary

  • Michael Tornaritis,

    Roles Investigation, Writing – review & editing

    Affiliation Research and Education Institute of Child Health, Strovolos, Cyprus

  • Licia Iacoviello,

    Roles Investigation, Writing – review & editing

    Affiliation Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo Neuromed, Pozzilli (IS), Italy

  • Yannis Pitsiladis,

    Roles Investigation, Writing – review & editing

    Affiliation University of Brighton, Eastbourne, United Kingdom

  • Claudio Galli †,

    † Deceased.

    Roles Investigation, Writing – review & editing

    Affiliation DiSFeB, Department of Pharmacological and Biomolecular Sciences, University of Milan, Milan, Italy

  • Ronja Foraita,

    Roles Methodology, Writing – review & editing

    Affiliation Leibniz Institute for Prevention Research and Epidemiology—BIPS, Bremen, Germany

  • Claudia Börnhorst,

    Roles Formal analysis, Methodology, Supervision

    Affiliation Leibniz Institute for Prevention Research and Epidemiology—BIPS, Bremen, Germany

  •  [ ... ],
  • on behalf of the IDEFICS and I. Family consortia
  • [ view all ]
  • [ view less ]

Abstract

Background

The recent obesity epidemic in children also showed an increase in the prevalence of hypertension. As blood pressure (BP) is associated with (long-chain) polyunsaturated fatty acids (LC PUFA), genetic variation in desaturase enzymes being involved in the synthesis of LC PUFA may be associated with BP. This study aimed to investigate the direct effects (independent of mediating variables) and indirect effects (mediated through intermediate variables) of a common variant in the FADS1 gene, rs174546, known to affect delta-5 desaturase (D5D) activity on PUFA level, body mass index (BMI) and BP.

Methods

A subsample of the IDEFICS (Identification and prevention of dietary- and lifestyle-induced health effects in children and infants) baseline survey including 520 children aged 2 to <10 years from six European countries was included. The association between rs174546 (T<C) and BP z-score as well as the mediating effects of selected key PUFA levels (dihomo-gamma-linolenic acid, DGLA; arachidonic acid, ARA; eicosapentaenoic acid, EPA) or estimated D5D activity (D5D index) and BMI z-score were investigated through path model analyses, adjusting for sex, age, educational level of parents, family history of hypertension, lifestyle factors and blood levels of saturated and monounsaturated fatty acids, triglycerides and low density lipoprotein cholesterol. Whole blood fatty acids were measured by a validated gas chromatographic method and recorded as percentage of weight of all fatty acids detected.

Results

Minor allele carriers of the SNP rs174546 had significantly higher DGLA and lower ARA and EPA levels as well as a lower D5D index. Via ARA and BMI z-score, the polymorphism had an indirect lowering effect on systolic BP z-score for each additional T allele (standardized effect estimate -0.057, p = 0.007). For DGLA, EPA and D5D index, the indirect effects of rs174546 on systolic BP were also negative but did not reach significance. DGLA and EPA had an increasing indirect effect on systolic BP via BMI. Results for diastolic BP were in general similar but effect estimates were lower compared to systolic BP.

Conclusion

Genetic variation in FADS1 influences BP via ARA and BMI indicating a favorable effect of the minor allele in SNP rs174546. Thus, polymorphisms with an impact on the D5D activity may play a role for the BP level mediated through PUFA and BMI. Therefore, health effects of dietary n-6 and n-3 PUFA may vary depending on genetic FADS1 variants.

Introduction

Blood pressure levels and the prevalence of hypertension are increasing in children and adolescents, particularly because of the epidemic of overweight and obesity already observed in young age groups [1, 2]. Blood and tissue levels of polyunsaturated fatty acids (PUFA), in particular long-chain (LC) PUFA have been shown to be associated with blood pressure (BP). Furthermore, circulating levels of n-3 LC PUFA were associated with a lower prevalence of hypertension [3, 4]. LC PUFA levels depend on dietary intake, bioavailability and PUFA metabolism. In the biosynthesis of LC PUFA from precursor PUFA, elongase and desaturase enzymes are crucial. FADS1, a member of the fatty acid (FA) desaturase gene cluster, codes for the enzyme delta-5 desaturase (D5D) which is involved in one step of the conversion of linoleic acid (LA, 18:2n-6) and alpha-linoleneic acid (ALA, 18:3n-3) to arachidonic acid (ARA, 20:4n-6) and eicosapentaenoic acid (EPA, 20:5n-3), respectively. Genetic variation in desaturases has shown to affect PUFA and LC PUFA tissue status and studies indicate that polymorphisms in this gene cluster also influence health outcomes like the lipoprotein profile and the risk of cardiovascular diseases [5]. Accordingly, estimated D5D activity (D5D index), calculated from the FA product/precursor ratio (20:4n-6/20:3n-6) was observed to be inversely associated with BMI [68] and BP [6, 9]. Thus, high D5D activity may be protective against adiposity and hypertension both of which are increasingly prevalent already in children and adolescents [10, 11] and tend to track into adulthood [1214].

Additionally, D5D activity strongly influences the PUFA status [15], particularly blood levels of EPA, docosahexaenoic acid (DHA, 22:6n-3) and ARA which have been linked to BP [3, 1618]. N-6 and n-3 LC PUFA can influence BP in several ways as they are precursors of prostaglandins, thromboxanes and oxidation products with BP modulating effects. While high ARA status has been shown to increase BP [4, 18, 19], n-3 LC PUFA seem to have BP lowering effects as indicated by several studies [3, 4, 20]. However, results of observational studies regarding n-3 LC PUFA and BP in children are inconsistent [18, 21, 22].

As polymorphisms in the FADS1 gene are associated with circulating PUFA levels [2327], the relation between LC PUFA and BP may be influenced by genetic variation. The single nucleotide polymorphism (SNP) rs174546 [24] was shown to have strong effects on serum PUFA levels. Carriers of the rs174546 minor (T) allele showed a reduced activity of D5D resulting in increased levels of the precursor n-6 PUFA LA and dihomo-gamma-linolenic acid (DGLA, 20:3n-6) as well as the n-3 precursor PUFA ALA while the n-6 and n-3 D5D products ARA and EPA were found to be decreased [24]. Several studies suggested that rs174546 plays an important functional role in the FADS1 gene cluster [24, 27, 28]. Results of the HELENA study investigating minor alleles of nine SNPs in the FADS gene cluster in a large group of adolescents suggest that the rs174546 SNP is in linkage disequilibrium with a functional SNP within FADS1 as all haplotypes carrying the minor allele were associated with lower D5D activity [24]. Also other studies indicated rs174546 to have a functional role in the FADS1 gene variants or to be at least an appropriate proxy for a functional but yet unknown FADS1 variant [27, 28]. Thus, we hypothesize that genetic variation in the FADS1 rs174546 influences BP via whole blood PUFA levels and BMI as mediators, and we investigate the direct and indirect effects on BP applying path analysis.

Methods

Study population

The IDEFICS (Identification and prevention of dietary- and lifestyle-induced health effects in children and infants) baseline survey (T0) included a population-based sample of 16,228 children aged 2 to 9.9 years from eight European countries (Belgium, Cyprus, Estonia, Germany, Hungary, Italy, Spain, Sweden) who were examined in 2007/2008. Follow-up examinations were conducted after two (T1) and six (T3, I.Family study) years; the design of this cohort study has been described in detail elsewhere [29, 30]. However, in this study, only cross-sectional data from the IDEFICS baseline survey were included as only at T0 a large number of children with FA data was available. In the IDEFICS study risk factors of overweight and obesity were investigated in young children and anthropometric and clinical examinations were conducted at each survey wave. Additionally, health characteristics and lifestyle behaviours were collected and biosamples were taken. Parents gave written informed consent prior to study participation and children gave oral consent prior to the examinations. All participating centers obtained ethical approval by the regional committees: Tallinn Medical Research Ethics Committee, Tallinn, Estonia; Ethics Committee of the University of Bremen, Bremen, Germany; Egészségügyi Tudományos Tanács, Pécs, Hungary; Azienda Sanitaria Locale Avellino Comitato Etico, Avellino, Italy; Regionala Etikprövningsnämnden i Göteborg, Gothenburg, Sweden; Comité Ético de Investigación Clínica de Aragón, Zaragoza, Spain.

Genotyping and quality control (QC) of SNP data

Genomic DNA was extracted from saliva samples. Genotyping was conducted using the UK Biobank Axiom® 96-Array from Affymetrix (Santa Clara, USA). Genotype calling was performed with respect to Affymetrix’s best practice guideline including analysis with SNPolisher assuming a dish QC > 0.82 and QC call rate > 0.97. 40 individuals were excluded for sample call rate ≤ 0.98, ambiguous sex assignments, heterozygosity or as population outliers assessed by principal component analysis [31, 32].

An additive effect was assumed, with 0, 1, or 2 encoding the number of minor allele T present at SNP rs174546. The analyses were limited to FADS1 as no genetic data on FADS2 was available.

Fatty acids

In a subsample of 2600 IDEFICS participants with an oversampling of overweight and obese children, FA profiles were analyzed [29]. Blood samples were obtained by collecting a drop of fasting blood from a fingertip or by venipuncture and were applied directly to an antioxidant-treated paper strip. FA in whole blood were separated and determined by a rapid gas-liquid chromatography method in a central laboratory as previously described [33, 34]. This method was validated by several laboratories [3436] and whole blood was shown to be representative for the total FA pool as it includes FA from all lipid classes with phospholipids of plasma and red blood cells (RBC) being the main contributors of the FA amount and profile [35]. Several studies confirmed that whole blood is a suitable marker for n-3 FA status and for long-term essential FA intake [34, 37, 38].

In the present analysis we used FA values of the n-3 PUFA, EPA and the n-6 PUFA, DGLA and ARA. FA are expressed as weight percentage of all FA detected (% wt/wt). Additionally, we calculated the D5D index from the product-precursor ratio (20:4n-6/20:3n-6).

Blood pressure

A standardized protocol was used to measure systolic (SBP) and diastolic (DBP) BP with an automated oscillometric device (Welch Allyn, Inc., 4200B-E2, Skaneateles Falls, NY, USA) after at least 5 minutes of rest in seated position at the right arm with a cuff of appropriate size for arm circumference [39]. Two recordings were taken, with a 2-min interval between each, plus a third measurement in case of a >5% difference between the first two readings. As outcome measure the mean value of the two measurements with the smallest difference was used. BP was set to missing if the first and second measurement deviated by >5% and no third measurement value was available. Age, sex and height specific reference values from the US Department of Health and Human Services, National Institutes of Health (NIH, 2005) [40] were applied to classify SBP, DBP and BP to normal BP, prehypertension and hypertension. Based on the NIH references age-, sex- and height-specific z-scores were calculated and used as outcome variables in the subsequent analysis.

BMI

Body weight was measured in light underwear in fasting state on a calibrated scale accurate to 0.1 kg (Tanita BC 420 MA Tanita Europe GmbH, Sindelfingen, Germany). A calibrated stadiometer (Seca 225 stadiometer, Birmingham, UK) was used to measure height of the children to the nearest 0.1 cm. BMI was calculated as weight [kg] divided by height [m] squared and BMI z-scores were calculated based on Cole & Lobstein (2012) [41].

Covariates.

Covariate data such as date of birth, dietary and lifestyle behaviors and socioeconomic information were assessed using questionnaires. In a personal interview the family history of diseases and the health conditions of the participating child were collected.

Sex of the child, age (continuous), country (six countries as categories: Estonia, Germany, Hungary, Italy, Spain, Sweden), maximum educational level of parents according to the International Standard Classification of Education (ISCED) [42] (three categories: low level, ISCED 0, 1 and 2; medium, ISCED 3 and 4; high, ISCED 5 and 6), sedentary behaviour (television and computer time (hours/week)), family history of hypertension (yes vs. no; as reported for biological parents/siblings) and birth weight (g) were considered as covariates. As indicator of dietary salt intake we used the weekly frequency of salty/snack food consumption which includes hamburger, hot dog, kebab, wrap, salty snacks such as savory pastries and fritters as well as cold cuts and ready to cook meat products. Further, the consumption frequency of fish (times/week) was considered in all analyses including EPA.

The weight percentages of our focus PUFA vary individually also depending on the proportions of the different lipid classes (phospholipids, triacylglycerols and cholesterylesters) in plasma, RBC, HDL and LDL. As total saturated (SFA) and total monounsaturated FA (MUFA) are the major FA with different distributions in these lipid classes [35], also weight percentages of SFA and MUFA of total FA were included as covariates. Their inclusion in the models aims to adjust for the different proportions of lipid classes. Additionally, we included blood levels of triglycerides (TG) and low-density lipoprotein cholesterol (LDL) as covariates. TG were directly measured with the ‘point-of-care analyser’ Cholestech LDX (Cholestech, Hayward, CA, USA) while LDL was calculated from Cholestech data as previously described [43].

Analysis group

Among the children with FA analyses (N = 2600), 10 children were excluded because of blood drawing in a non-fasting state and further 406 children had to be excluded due to missing BP or missing covariate data resulting in 2184 children. No genetic data was available for 1664 children, leaving 520 children in the final analysis sample (Fig 1).

thumbnail
Fig 1. Flow chart of the inclusion and exclusion of IDEFICS participants.

https://doi.org/10.1371/journal.pone.0181485.g001

Statistical analyses

Path analysis was used to test a theory-driven model linking FADS1 to blood pressure mediated by effects on FA and BMI. Path analysis has the advantage over linear regression that path coefficients are estimated via simultaneous equation estimation, thus adjusting standard errors for the number of equations computed [44]. These models further allow the estimation of total, direct and indirect effects. The structure of our theoretical model is displayed in Fig 2. The arrows in Fig 2 represent direct effects which are independent of the other variables in the path model. Additionally, there are indirect effects which influence a variable via an intermediate variable (not presented in Fig 2) [45]. This means that FADS1 was assumed to affect BP directly but also indirectly through FA status (or D5D index) and BMI that lie on paths between FADS1 and BP. The FA status (or D5D index) was assumed to affect BP directly as well as via BMI and the BMI was assumed to directly affect BP. Separate models were run for the two outcomes SBP and DBP as well as for the four exposures (1) DGLA, (2) ARA, (3) EPA and (4) D5D. Models 1a to 4a refer to the four models run for SBP, models 1b to 4b to the four models run for DBP, respectively. The covariates mentioned above were considered in the analyses (see S1 Table for a more detailed description of the covariates being related to the different endogenous (outcome) variables in the model). After running the a priori defined path model, the following model modifications were applied based on both theoretical considerations and model fit indices: covariances among residual terms, among country indicators and sex, among sex and age, sex and educational level of parents, educational level of parents and age, educational level of parents and SNP genotypes, SNP genotypes and age, SNP genotypes and media time as well as birth weight and media time were set to zero. Based on fit indices, also some changes in the covariates affecting the three endogenous variables BP, BMI and FA were applied if biologically plausible as summarized in S1 Table. For parameter estimation, we used the maximum-likelihood procedure with quasi-Newton optimization. A non-significant Chi-square test (P>0.05), Comparative Fit Index (CFI) ≥0.97, and Root Mean Square Error of Approximation (RMSEA) ≤0.03 indicated a very good model fit in all models estimated [46].

thumbnail
Fig 2. Theoretical models displaying the direct effects in proposed pathways through which BP (SBP or DBP, respectively) is affected by the SNP (rs174546), by fatty acids (DGLA, ARA, EPA, D5D index) as well as by BMI.

https://doi.org/10.1371/journal.pone.0181485.g002

Mediation was assessed by computing estimates of direct, indirect, and total effects of the associations specified in the model. Direct effects represent associations between variables unmediated by any other variable in the model. Indirect effects represent mediated effects (or combined mediated effects for paths through multiple mediators). Total effects are the sum of the direct and indirect effects. For all models, standardized path coefficients and corresponding p-values are reported. Following the proposal by Cribbie (2007) [47] of how to account for multiple testing problems in pathway analysis, the adjustment method by Benjamini & Hochberg (1995) [48] was applied to control for the false discovery rate (FDR) at the 0.05 level of significance. Multiple comparison adjustment was applied for each outcome (SBP and DBP) and each research question resulting in different adjusted alpha levels which are indicated in the respective result table. All analyses were performed using SAS® statistical software version 9.3 (SAS Institute, Inc., Cary, NC). All path models were run using SAS Proc CALIS.

Results

Table 1 gives a description of the total study group by blood pressure category. Children classified as hypertensive had a higher mean BMI z-score, a lower mean birth weight and a higher mean consumption frequency of salty/snack food compared with normotensive children. Table 2 shows the characteristics of the study population by rs174546 genotypes. The SNP was in Hardy-Weinberg Equilibrium (p = 0.395) and minor allele frequency for rs174546 was 29%. About 9% of children were homozygous and about 40% were heterozygous carriers of the minor allele.

thumbnail
Table 1. Baseline characteristics of the study population by blood pressure category[1].

https://doi.org/10.1371/journal.pone.0181485.t001

thumbnail
Table 2. Baseline characteristics of the study population by FADS1 genotype.

https://doi.org/10.1371/journal.pone.0181485.t002

Pairwise correlations among the SNP, PUFA, BMI z-score, SBP z-score and DBP z-score are displayed in Table 3. Significant positive correlations were observed between BMI z-score and SBP z-score, DGLA and ARA. DGLA was positively correlated with the SNP. In contrast, the number of T alleles at SNP rs174546 correlated inversely with ARA and D5D index.

thumbnail
Table 3. Pearson correlation coefficients and p-values among SBP, DBP, BMI z-score, FA and D5D index and the SNP.

https://doi.org/10.1371/journal.pone.0181485.t003

Table 4 shows the estimated direct and indirect effects obtained from the different path models to disentangle the associations among BP z-score, BMI z-score, the FA and D5D index as well as the SNP rs174546. Models 1a to 4a present the results obtained for SBP with regard to DGLA (model 1a), ARA (model 1b), EPA (model 1c) and D5D (model 1d), and models 1b to 4b present the respective results for DBP. Standardized direct as well as indirect effects are displayed according to the defined path model. As expected, rs174546 showed significant direct effects on all FA levels and D5D index being positive for DGLA and negative for ARA, EPA and D5D index. The direct effects of the SNP on BMI were negative in all models but reached statistical significance only in the models for DGLA and for the D5D index. Neither the SNP nor the FA or D5D index showed a direct effect on SBP or DBP in any of the models. A direct effect on SBP and DBP was only observed for the BMI z-score. DGLA and EPA showed a positive direct effect on BMI.

thumbnail
Table 4. Standardized estimates of direct and indirect effects and corresponding p-values obtained from path analyses for the associations among systolic/diastolic blood pressure z-score, BMI z-score, the FA and D5D index and the SNP rs174546.

https://doi.org/10.1371/journal.pone.0181485.t004

SNP rs174546 revealed a positive indirect effect on BMI via DGLA. A negative indirect effect of the SNP on SBP via ARA and BMI was observed. Fig 3 presents the estimated direct effects which are independent of mediating variables between the SNP rs174546, ARA, BMI and SBP. The largest effect observed was between BMI and SBP indicating that a higher BMI was associated with an increase in SBP. Further, there was a strong negative effect of the SNP on ARA indicating that in minor allele carriers the conversion from DGLA to ARA is reduced. As can be seen from Fig 3, the negative indirect effect between the SNP and SBP results from the negative direct effects of the SNP on ARA as well as on BMI overlaying the positive direct effects of ARA and the BMI on SBP.

thumbnail
Fig 3. Path model for the associations between the SNP rs174546 and SBP z-score via ARA and BMI z-score: Standardized direct effects and p-values.

https://doi.org/10.1371/journal.pone.0181485.g003

DGLA and EPA had a positive indirect effect on SBP via BMI. The observed indirect effects on SBP point to the same directions for DBP.

The positive direct effect of EPA on BMI was unexpected and previous analyses indicated that the association of the EPA status with BP may depend on weight status [18]. Therefore, an additional analysis was conducted assessing the direct effects of EPA on BMI z-score and BP z-scores and the indirect effects of EPA on BP z-scores through BMI z-score separately for thin/normal weight versus overweight/obese children. This was done by use of respective interaction terms between weight status and EPA in the path model. The results are displayed in Table 5. EPA had a direct negative effect on BMI in thin/normal weight children whereas a positive effect was observed in overweight/obese children. The direct and indirect effects of EPA on BP point to the same directions, i.e. were negative in thin/normal weight children and positive in overweight/obese children.

thumbnail
Table 5. Standardized estimates of direct effects of EPA on BMI z-score and SBP z-score/DBP z-score and indirect effects of EPA on SBP z-score/DBP z-score via BMI z-score in thin/normal weight versus overweight/obese children obtained from path analyses.

https://doi.org/10.1371/journal.pone.0181485.t005

Discussion

Our study in European children revealed that carrying the minor (T) variant of SNP rs174546, lying in the FADS1 gene, has a favorable indirect effect on SBP via ARA in an analysis based on path models. However, no direct effect of the T allele on BP was observed. This indirect effect results from various countervailing direct effects via the blood level of ARA and the BMI z-score. The SNP was associated with lower ARA levels and this association may have overlaid the positive associations of ARA and the BMI with SBP in our model. To our knowledge this is the first study showing that the investigated FADS1 SNP has a beneficial effect on SBP via its negative effects on the ARA level and on BMI z-score. This is in line with previous findings indicating that high ARA is associated with elevated BP in adults [4, 19] and, as observed in our recent analysis, also in children [18]. As expected, we observed strong negative effects of the SNP on ARA because in minor allele carriers the D5D-mediated conversion from DGLA to ARA is reduced. Further, the BMI z-score showed the expected positive association with BP.

While we observed a beneficial indirect effect of the SNP via ARA and the BMI on SBP, the direct association of the SNP with SBP indicated a positive direction which was not statistically significant (0.077, p = 0.085; total effect: 0.020, p = 0.641). Also the direct effects of ARA on BMI and the direct effects of ARA as well as BMI on SBP pointed to a positive direction although only the latter was statistically significant. The opposite direct effect of the SNP on SBP may result from factors not considered in the respective model which could have been influenced by the SNP, like the levels of other FA, e.g. decreased DHA [26] or increased insulin levels [6] in minor allele carriers, both of which can contribute to BP elevation [3, 16, 49].

For DGLA, EPA and D5D index, the indirect effects also pointed to a negative direction indicating a beneficial effect of the SNP though not being statistically significant. For DGLA, a beneficial indirect effect of the SNP on BP would be unlikely as more DGLA is available in minor allele carriers and DGLA had a positive indirect effect on BP in our study via BMI and was positively associated with BP in previous studies [17, 50, 51]. Additionally, high DGLA levels result in a lower D5D index which is the ratio of ARA and DGLA. The D5D index has shown to be inversely associated with BP in adults [6, 52]. Therefore, one may hypothesize that there is a positive association of rs174546 with BP via D5D index as minor allele carriers had a lower D5D index. However, because of the expected and observed strong negative direct effect of the SNP on the D5D index together with its negative direct effect on BMI, the resulting indirect effect of the SNP on SBP via D5D index and BMI pointed to a negative direction.

Additionally, our results suggest a positive indirect effect of the SNP on BMI via DGLA. The latter result is in line with a study in adolescents indicating a positive association of DGLA and BMI [53].

Contrary to our expectations, there was no significant indirect effect of the common FADS1 SNP on BP via EPA and BMI. The SNP was assumed to be positively associated with BP via EPA as in children with minor alleles, EPA biosynthesis is reduced. High EPA levels were shown to be protective against hypertension in previous studies in adults [3, 4] although results in children are inconsistent [18, 21, 22]. In our model, EPA unexpectedly showed a direct positive effect on BMI and an indirect positive effect on SBP via BMI. However, a subgroup analysis revealed that the direct effect of EPA on BMI was only present in overweight/obese children whereas in thin/normal weight children the expected negative effect was observed. The indirect effect of EPA on BP also pointed to a positive effect in overweight/obese children but to a negative effect in thin/normal weight children. Thus, the unexpected effects of EPA are likely to result from effect modification by weight status. Also in a recent study, we found a beneficial effect of EPA on BP only in thin/normal weight children whereas an unfavorable association was observed in overweight/obese children [18]. We did not consider DHA in our analyses as the conversion efficiency from ALA and EPA to DHA is very low [54, 55] and we were particularly interested in the SNP’s influence on endogenous synthesis via FADS1 and the effects on BP in this analysis. Additionally, in previous analyses no statistically significant effect of DHA on BP was observed neither in the total sample nor in an analysis stratified by weight status [18].

As reported in previous studies [2327], our results confirm for the n-6 PUFA that the SNP directly affects the FA levels indicating that minor allele carriers had higher levels of precursor (DGLA) and lower levels of product FA (ARA) whereas for n-3 PUFA this effect is less obvious. This was also reflected by a negative effect on the D5D index in our study.

The rs174546 SNP has also shown to influence lipoprotein profiles as T allele carriers have lower cholesterol levels [56, 57]. Additionally, LDL and TG have been shown to influence the association between BMI and BP [58] and were hence included as covariates in the models. FADS1 gene variants and cardiovascular diseases (CVD) or stroke for which hypertension is a major risk factor suggested a role of the rs174546 SNP. In a Chinese case-control study, the rs174546 T allele was found to be associated with an increased risk of coronary artery disease and ischemic stroke while the CC/CT genotypes were associated with a decreased risk [28]. In a large Danish study including more than 24,032 participants with 2648 CVD cases during 14 years of follow-up, no effect of rs174546 on CVD risk in the total study group was observed. However, among participants homozygous for the minor T allele of rs174546, subjects with a higher dietary intake of the n-3 FA ALA and a higher ratio of ALA to the n-6 FA LA tended to be at lower risk for CVD and ischemic stroke [59]. This may indicate a protective effect of the resulting higher EPA/ARA ratio. In our study the SNP had a negative effect on SBP via lower ARA. The observed trend for a negative effect of the SNP on SBP via EPA in the total study group depended on the weight status and indicated a protective role of EPA only in thin/normal weight children. The effects of variants of a common FADS1 SNP on BP as shown in our study are one element that has to be considered in the complex association of n-6 and n-3 LC PUFA with cardiovascular risk factors.

Limitations and strengths

Intrinsic variability of BP could not be considered in our study because repeated BP measurements could be taken only on a single occasion on the examination day. This may have led to an overestimation of BP values [40, 60]. As we had no reliable information on dietary sodium intake, we used the intake frequency of salty/snack foods as a proxy for sodium intake. These and other processed foods have shown to be the main source of salt [61]. As absolute quantitative data of FA were not available we included weight percentages of SFA and MUFA of total FA as covariates to adjust for the different proportions of lipid classes.

As all data were measured cross-sectionally in the present study, thus causality cannot be determined.

Important strengths of our study are the pan-European analysis sample of 520 children as well as the detailed phenotyping of the children. In all countries standardized protocols for data assessment, measurements and biosampling were applied and adherence to protocols was verified by central quality control. To investigate the association of the SNP and BP in children can be an advantage as the genetic basis may be stronger compared to adults because of the accumulation of behavioral and lifetime exposures in older indviduals [62, 63].

Conclusions

Our results indicate that genetic variation in the FADS1 gene, rs174546, influences BP via ARA and BMI. Thus, polymorphisms with an impact on the D5D activity may play a role for the BP level mediated through PUFA and BMI. Therefore, health effects of dietary n-6 and n-3 PUFA may vary depending on genetic FADS1 variants and—considering the observed effect modification—on weight status. Different dietary intake recommendations for LC PUFA may be required according to genotype and weight status.

Supporting information

S1 Table. Exposures and covariates selected a priori for the three endogenous variables (BP, BMI z-score and FA and D5D index) as well as after model modification based on fit indices and theoretical considerations (final model)

https://doi.org/10.1371/journal.pone.0181485.s001

(DOCX)

Acknowledgments

We wish to thank the IDEFICS/I.Family children and their parents for participating in these extensive examinations.

The whole blood fatty acids in IDEFICS biosamples were analysed by the team at the University of Milan under the lead of Prof. Claudio Galli, who sadly passed away on 11 April 2017. He contributed substantially to the development of this and other manuscripts on fatty acids by giving wise advice and encouraging support, for which we are most thankful.

References

  1. 1. Martin L, Oepen J, Reinehr T, Wabitsch M, Claussnitzer G, Waldeck E, et al. Ethnicity and cardiovascular risk factors: evaluation of 40,921 normal-weight, overweight or obese children and adolescents living in Central Europe. Int J Obes (Lond). 2015;39(1):45–51. Epub 2014/09/13. pmid:25214149.
  2. 2. de Moraes AC, Carvalho HB, Siani A, Barba G, Veidebaum T, Tornaritis M, et al. Incidence of high blood pressure in children—effects of physical activity and sedentary behaviors: the IDEFICS study: High blood pressure, lifestyle and children. Int J Cardiol. 2015;180:165–70. Epub 2014/12/03. pmid:25460372.
  3. 3. Zeng FF, Sun LL, Liu YH, Xu Y, Guan K, Ling WH, et al. Higher erythrocyte n-3 PUFAs are associated with decreased blood pressure in middle-aged and elderly Chinese adults. J Nutr. 2014;144(8):1240–6. Epub 2014/06/27. pmid:24966412.
  4. 4. Zheng ZJ, Folsom AR, Ma J, Arnett DK, McGovern PG, Eckfeldt JH. Plasma fatty acid composition and 6-year incidence of hypertension in middle-aged adults: the Atherosclerosis Risk in Communities (ARIC) Study. Am J Epidemiol. 1999;150(5):492–500. Epub 1999/09/03. pmid:10472949.
  5. 5. Merino DM, Ma DW, Mutch DM. Genetic variation in lipid desaturases and its impact on the development of human disease. Lipids Health Dis. 2010;9:63. Epub 2010/06/23. pmid:20565855; PubMed Central PMCID: PMCPmc2914715.
  6. 6. Murakami K, Sasaki S, Takahashi Y, Uenishi K, Watanabe T, Kohri T, et al. Lower estimates of delta-5 desaturase and elongase activity are related to adverse profiles for several metabolic risk factors in young Japanese women. Nutr Res. 2008;28(12):816–24. Epub 2008/12/17. pmid:19083494.
  7. 7. Abe Y, Okada T, Iguchi H, Saito E, Kuromori Y, Iwata F, et al. Association of changes in body fatness and fatty acid composition of plasma phospholipids during early puberty in Japanese children. J Atheroscler Thromb. 2012;19(12):1102–9. DN/JST.JSTAGE/jat/12476 [pii]. pmid:22863834
  8. 8. Wolters M, Schlenz H, Bornhorst C, Rise P, Galli C, Moreno LA, et al. Desaturase activity is associated with weight status and metabolic risk markers in young children. J Clin Endocrinol Metab. 2015;100(10):3760–9. Epub 2015/08/19. pmid:26284759.
  9. 9. Kim SR, Jeon SY, Lee SM. The association of cardiovascular risk factors with saturated fatty acids and fatty acid desaturase indices in erythrocyte in middle-aged Korean adults. Lipids Health Dis. 2015;14:133. Epub 2015/10/27. pmid:26497880; PubMed Central PMCID: PMCPmc4619393.
  10. 10. Xi B, Zhang T, Zhang M, Liu F, Zong X, Zhao M, et al. Trends in Elevated Blood Pressure Among US Children and Adolescents: 1999–2012. Am J Hypertens. 2016;29(2):217–25. Epub 2015/07/15. pmid:26158854.
  11. 11. de Moraes AC, Lacerda MB, Moreno LA, Horta BL, Carvalho HB. Prevalence of high blood pressure in 122,053 adolescents: a systematic review and meta-regression. Medicine (Baltimore). 2014;93(27):e232. Epub 2014/12/17. pmid:25501086; PubMed Central PMCID: PMCPmc4602805.
  12. 12. Liang Y, Mi J. Pubertal hypertension is a strong predictor for the risk of adult hypertension. Biomed Environ Sci. 2011;24(5):459–66. Epub 2011/11/24. pmid:22108410.
  13. 13. Kollias A, Pantsiotou K, Karpettas N, Roussias L, Stergiou GS. Tracking of blood pressure from childhood to adolescence in a Greek cohort. Eur J Public Health. 2012;22(3):389–93. Epub 2011/06/28. pmid:21705785.
  14. 14. Singh AS, Mulder C, Twisk JW, van Mechelen W, Chinapaw MJ. Tracking of childhood overweight into adulthood: a systematic review of the literature. Obes Rev. 2008;9(5):474–88. Epub 2008/03/12. pmid:18331423.
  15. 15. Tosi F, Sartori F, Guarini P, Olivieri O, Martinelli N. Delta-5 and delta-6 desaturases: crucial enzymes in polyunsaturated fatty acid-related pathways with pleiotropic influences in health and disease. Adv Exp Med Biol. 2014;824:61–81. Epub 2014/07/21. pmid:25038994.
  16. 16. Virtanen JK, Nyantika AN, Kauhanen J, Voutilainen S, Tuomainen TP. Serum long-chain n-3 polyunsaturated fatty acids, methylmercury and blood pressure in an older population. Hypertens Res. 2012;35(10):1000–4. Epub 2012/06/08. pmid:22673531.
  17. 17. Ebbesson SO, Voruganti VS, Higgins PB, Fabsitz RR, Ebbesson LO, Laston S, et al. Fatty acids linked to cardiovascular mortality are associated with risk factors. Int J Circumpolar Health. 2015;74:28055. Epub 2015/08/15. pmid:26274054.
  18. 18. Wolters M, Pala V, Russo P, Rise P, Moreno LA, De Henauw S, et al. Associations of whole blood n-3 and n-6 polyunsaturated fatty acids with blood pressure in children and adolescents–results from the IDEFICS/I.Family cohort. PLoS One. 2016;11(11):1–17. pmid:27806134
  19. 19. Williams ES, Baylin A, Campos H. Adipose tissue arachidonic acid and the metabolic syndrome in Costa Rican adults. Clin Nutr. 2007;26(4):474–82. Epub 2007/05/18. pmid:17507118; PubMed Central PMCID: PMCPmc2730166.
  20. 20. Skilton MR, Raitakari OT, Celermajer DS. High intake of dietary long-chain omega-3 fatty acids is associated with lower blood pressure in children born with low birth weight: NHANES 2003–2008. Hypertension. 2013;61(5):972–6. Epub 2013/03/06. pmid:23460284.
  21. 21. Damsgaard CT, Eidner MB, Stark KD, Hjorth MF, Sjodin A, Andersen MR, et al. Eicosapentaenoic acid and docosahexaenoic acid in whole blood are differentially and sex-specifically associated with cardiometabolic risk markers in 8-11-year-old danish children. PLoS One. 2014;9(10):e109368. Epub 2014/10/21. pmid:25330302; PubMed Central PMCID: PMCPmc4198100.
  22. 22. Lauritzen L, Harslof LB, Hellgren LI, Pedersen MH, Molgaard C, Michaelsen KF. Fish intake, erythrocyte n-3 fatty acid status and metabolic health in Danish adolescent girls and boys. Br J Nutr. 2012;107(5):697–704. pmid:21736784.
  23. 23. Schuchardt JP, Kobe T, Witte V, Willers J, Gingrich A, Tesky V, et al. Genetic Variants of the FADS Gene Cluster Are Associated with Erythrocyte Membrane LC PUFA Levels in Patients with Mild Cognitive Impairment. J Nutr Health Aging. 2016;20(6):611–20. Epub 2016/06/09. pmid:27273350.
  24. 24. Bokor S, Dumont J, Spinneker A, Gonzalez-Gross M, Nova E, Widhalm K, et al. Single nucleotide polymorphisms in the FADS gene cluster are associated with delta-5 and delta-6 desaturase activities estimated by serum fatty acid ratios. J Lipid Res. 2010;51(8):2325–33. Epub 2010/04/30. pmid:20427696; PubMed Central PMCID: PMCPmc2903808.
  25. 25. Mathias RA, Vergara C, Gao L, Rafaels N, Hand T, Campbell M, et al. FADS genetic variants and omega-6 polyunsaturated fatty acid metabolism in a homogeneous island population. J Lipid Res. 2010;51(9):2766–74. Epub 2010/06/22. pmid:20562440; PubMed Central PMCID: PMCPmc2918459.
  26. 26. Koletzko B, Lattka E, Zeilinger S, Illig T, Steer C. Genetic variants of the fatty acid desaturase gene cluster predict amounts of red blood cell docosahexaenoic and other polyunsaturated fatty acids in pregnant women: findings from the Avon Longitudinal Study of Parents and Children. Am J Clin Nutr. 2011;93(1):211–9. Epub 2010/11/26. pmid:21106917.
  27. 27. Zietemann V, Kroger J, Enzenbach C, Jansen E, Fritsche A, Weikert C, et al. Genetic variation of the FADS1 FADS2 gene cluster and n-6 PUFA composition in erythrocyte membranes in the European Prospective Investigation into Cancer and Nutrition-Potsdam study. Br J Nutr. 2010;104(12):1748–59. Epub 2010/08/10. pmid:20691134.
  28. 28. Yang Q, Yin RX, Cao XL, Wu DF, Chen WX, Zhou YJ. Association of two polymorphisms in the FADS1/FADS2 gene cluster and the risk of coronary artery disease and ischemic stroke. Int J Clin Exp Pathol. 2015;8(6):7318–31. Epub 2015/08/12. pmid:26261632; PubMed Central PMCID: PMCPmc4525966.
  29. 29. Ahrens W, Bammann K, Siani A, Buchecker K, De HS, Iacoviello L, et al. The IDEFICS cohort: design, characteristics and participation in the baseline survey. Int J Obes (Lond). 2011;35 Suppl 1:S3–15. pmid:21483420
  30. 30. Ahrens W, Siani A, Adan R, De Henauw S, Eiben G, Gwozdz W, et al. Cohort Profile: The transition from childhood to adolescence in European children-how I.Family extends the IDEFICS cohort. Int J Epidemiol. 2016. Epub 2017/01/04. pmid:28040744.
  31. 31. Weale ME. Quality control for genome-wide association studies. Methods Mol Biol. 2010;628:341–72. Epub 2010/03/20. pmid:20238091.
  32. 32. Ziegler A. Genome-wide association studies: quality control and population-based measures. Genet Epidemiol. 2009;33 Suppl 1:S45–50. Epub 2009/11/20. pmid:19924716; PubMed Central PMCID: PMCPmc2996103.
  33. 33. Wolters M, Schlenz H, Foraita R, Galli C, Rise P, Moreno LA, et al. Reference values of whole-blood fatty acids by age and sex from European children aged 3–8 years. Int J Obes (Lond). 2014;38 Suppl 2:S86–98. pmid:25219413; PubMed Central PMCID: PMC4165865.
  34. 34. Marangoni F, Colombo C, Galli C. A method for the direct evaluation of the fatty acid status in a drop of blood from a fingertip in humans: applicability to nutritional and epidemiological studies. Anal Biochem. 2004;326(2):267–72. pmid:15003567
  35. 35. Rise P, Eligini S, Ghezzi S, Colli S, Galli C. Fatty acid composition of plasma, blood cells and whole blood: relevance for the assessment of the fatty acid status in humans. Prostaglandins Leukot Essent Fatty Acids. 2007;76(6):363–9. Epub 2007/06/26. pmid:17588731.
  36. 36. Rise P, Tragni E, Ghezzi S, Agostoni C, Marangoni F, Poli A, et al. Different patterns characterize Omega 6 and Omega 3 long chain polyunsaturated fatty acid levels in blood from Italian infants, children, adults and elderly. Prostaglandins Leukot Essent Fatty Acids. 2013;89(4):215–20. pmid:23910046
  37. 37. Harris WS, Pottala JV, Varvel SA, Borowski JJ, Ward JN, McConnell JP. Erythrocyte omega-3 fatty acids increase and linoleic acid decreases with age: observations from 160,000 patients. Prostaglandins Leukot Essent Fatty Acids. 2013;88(4):257–63. pmid:23375840
  38. 38. Baylin A, Kim MK, Donovan-Palmer A, Siles X, Dougherty L, Tocco P, et al. Fasting whole blood as a biomarker of essential fatty acid intake in epidemiologic studies: comparison with adipose tissue and plasma. Am J Epidemiol. 2005;162(4):373–81. pmid:16014782
  39. 39. Alpert BS. Validation of the Welch Allyn Spot Vital Signs blood pressure device according to the ANSI/AAMI SP10: 2002. Accuracy and cost-efficiency successfully combined. Blood Press Monit. 2007;12(5):345–7. Epub 2007/09/25. pmid:17890975.
  40. 40. National Institutes of Health. The fourth report on the diagnosis, evaluation, and treatment of high blood pressure in children and adolescents. NIH Publication No. 05–5267. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES, May 2005, Originally printed September 1996 (96–3790) Contract No.:
  41. 41. Cole TJ, Lobstein T. Extended international (IOTF) body mass index cut-offs for thinness, overweight and obesity. Pediatr Obes. 2012;7(4):284–94. pmid:22715120
  42. 42. United Nations. International Standard Classification of Education. http://www.uis.unesco.org/Library/Documents/isced97-en.pdf (accessed May 2016), 1997, revised 2006.
  43. 43. De Henauw S, Michels N, Vyncke K, Hebestreit A, Russo P, Intemann T, et al. Blood lipids among young children in Europe: results from the European IDEFICS study. Int J Obes (Lond). 2014;38 Suppl 2:S67–75. pmid:25219412.
  44. 44. Bollen KA. Structural Equations with Latent Variables. New York: Wiley-Interscience; 1989.
  45. 45. Stage FK, Carter HC, Nora A. Path Analysis: An Introduction and Analysis of a Decade of Research. The Journal of Educational Research. 2004;98(1):5–13.
  46. 46. Kline RB. Principles and Practice of Structural Equation Modeling. New York: The Guilford Press; 2011.
  47. 47. Cribbie RA. Multiplicity Control in Structural Equation Modeling. Structural Equation Modeling: A Multidisciplinary Journal. 2007;14(1):98–112.
  48. 48. Benjamini Y, Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society Series B (Methodological). 1995;57(1):289–300.
  49. 49. Reaven GM. Insulin resistance: the link between obesity and cardiovascular disease. Med Clin North Am. 2011;95(5):875–92. Epub 2011/08/23. pmid:21855697.
  50. 50. Simon JA, Fong J, Bernert JT Jr. Serum fatty acids and blood pressure. Hypertension. 1996;27(2):303–7. Epub 1996/02/01. pmid:8567056.
  51. 51. Grimsgaard S, Bonaa KH, Jacobsen BK, Bjerve KS. Plasma saturated and linoleic fatty acids are independently associated with blood pressure. Hypertension. 1999;34(3):478–83. Epub 1999/09/18. pmid:10489397.
  52. 52. Yang B, Ding F, Yan J, Ye XW, Xu XL, Wang FL, et al. Exploratory serum fatty acid patterns associated with blood pressure in community-dwelling middle-aged and elderly Chinese. Lipids Health Dis. 2016;15(1):58. Epub 2016/03/20. pmid:26993260.
  53. 53. Steffen LM, Vessby B, Jacobs DR Jr., Steinberger J, Moran A, Hong CP, et al. Serum phospholipid and cholesteryl ester fatty acids and estimated desaturase activities are related to overweight and cardiovascular risk factors in adolescents. Int J Obes (Lond). 2008;32(8):1297–304. pmid:18560369; PubMed Central PMCID: PMC2832613.
  54. 54. Brenna JT, Salem N Jr., Sinclair AJ, Cunnane SC. alpha-Linolenic acid supplementation and conversion to n-3 long-chain polyunsaturated fatty acids in humans. Prostaglandins Leukot Essent Fatty Acids. 2009;80(2–3):85–91. Epub 2009/03/10. pmid:19269799.
  55. 55. Burdge GC, Calder PC. Conversion of alpha-linolenic acid to longer-chain polyunsaturated fatty acids in human adults. Reprod Nutr Dev. 2005;45(5):581–97. Epub 2005/09/29. pmid:16188209.
  56. 56. Sone Y, Kido T, Ainuki T, Sonoda M, Ichi I, Kodama S, et al. Genetic variants of the fatty acid desaturase gene cluster are associated with plasma LDL cholesterol levels in Japanese males. J Nutr Sci Vitaminol (Tokyo). 2013;59(4):325–35. Epub 2013/09/26. pmid:24064733.
  57. 57. Molto-Puigmarti C, Jansen E, Heinrich J, Standl M, Mensink RP, Plat J, et al. Genetic variation in FADS genes and plasma cholesterol levels in 2-year-old infants: KOALA Birth Cohort Study. PLoS One. 2013;8(5):e61671. Epub 2013/05/15. pmid:23667444; PubMed Central PMCID: PMCPmc3648514.
  58. 58. Zou ZY, Yang YD, Wang S, Dong B, Li XH, Ma J. The importance of blood lipids in the association between BMI and blood pressure among Chinese overweight and obese children. Br J Nutr. 2016;116(1):45–51. Epub 2016/05/14. pmid:27169905.
  59. 59. Hellstrand S, Ericson U, Gullberg B, Hedblad B, Orho-Melander M, Sonestedt E. Genetic variation in FADS1 has little effect on the association between dietary PUFA intake and cardiovascular disease. J Nutr. 2014;144(9):1356–63. Epub 2014/07/11. pmid:25008580; PubMed Central PMCID: PMCPmc4130826.
  60. 60. Barba G, Buck C, Bammann K, Hadjigeorgiou C, Hebestreit A, Marild S, et al. Blood pressure reference values for European non-overweight school children: the IDEFICS study. Int J Obes (Lond). 2014;38 Suppl 2:S48–56. Epub 2014/09/16. pmid:25219411.
  61. 61. Marrero NM, He FJ, Whincup P, Macgregor GA. Salt intake of children and adolescents in South London: consumption levels and dietary sources. Hypertension. 2014;63(5):1026–32. Epub 2014/03/13. pmid:24614217.
  62. 62. Yang HC, Liang YJ, Wu YL, Chung CM, Chiang KM, Ho HY, et al. Genome-wide association study of young-onset hypertension in the Han Chinese population of Taiwan. PLoS One. 2009;4(5):e5459. Epub 2009/05/08. pmid:19421330; PubMed Central PMCID: PMCPMC2674219.
  63. 63. Simino J, Shi G, Bis JC, Chasman DI, Ehret GB, Gu X, et al. Gene-age interactions in blood pressure regulation: a large-scale investigation with the CHARGE, Global BPgen, and ICBP Consortia. Am J Hum Genet. 2014;95(1):24–38. Epub 2014/06/24. pmid:24954895; PubMed Central PMCID: PMCPMC4085636.