Sex Differences in Blood HDL-c, the Total Cholesterol/HDL-c Ratio, and Palmitoleic Acid are Not Associated with Variants in Common Candidate Genes

Abstract

Blood lipids are associated with cardiovascular disease (CVD) risk. Moreover, circulating lipid and fatty acid levels vary between men and women, and evidence demonstrates these traits may be influenced by single nucleotide polymorphisms (SNP). Sex-genotype interactions related to blood lipids and fatty acids have been poorly investigated and may help elucidate sex differences in CVD risk. The goal of this study was to investigate if the influence of SNPs previously associated with blood lipids and fatty acids varies in a sex-specific manner. Lipids and fatty acids were measured in serum and red blood cells (RBC), respectively, in 94 adults (18–30 years) from the GONE FISHIN’ cohort and 118 age-matched individuals from the GOLDN cohort. HDL-c levels were higher and the total cholesterol/HDL-c (TC/HDL-c) ratio was lower in women versus men (p < 0.01). RBC palmitoleic acid and the stearoyl-CoA desaturase index were both higher in women (p < 0.01). Fatty acid desaturase (FADS) pathway activity (estimated using the ratio of eicosapentaenoic acid/alpha-linolenic acid) was higher in men (p < 0.01). The AA genotype for rs1800775 in CETP had a lower TC/HDL-c ratio in men, but not women (p_int = 0.03). Independent of sex, major alleles for rs174537 in FADS1 (GG) and rs3211956 in CD36 (TT) had higher arachidonic acid, lower dihomo-γ-linoleic acid, and a higher FADS1 activity compared to minor alleles. The current study showed that blood lipid and fatty acid levels vary between healthy young men and women, but that the observed sex differences are not associated with common variants in candidate lipid metabolism genes.

Introduction

Sex-specific differences in cellular, metabolic, and physiological processes are well recognized and thought to contribute directly to differences in cardiovascular disease (CVD) risk between men and women [1]. Dyslipidemia represents an important determinant of CVD risk [2], and was previously reported to have a stronger association with risk of acute myocardial infarction in men compared to women [3]. Broadly speaking, both blood lipid (e.g., low-density lipoprotein cholesterol, LDL-c; high-density lipoprotein cholesterol, HDL-c; triacylglycerol, TAG) and fatty acid (FA; saturated fatty acids, SFA; monounsaturated fatty acids, MUFA; polyunsaturated fatty acids, PUFA) profiles have been associated with CVD risk and shown to differ between sexes [4, 5]. For example, pre-menopausal women have higher HDL-c levels compared to men [6], while men tend to have higher TAG and LDL-c levels compared to women [7–9]. Similarly, sex differences have also been reported in the FA composition of blood phospholipids. For example, α-linolenic acid (ALA) and docosahexaenoic acid (DHA) are on average higher in women compared to men [10], while the levels of SFA and MUFA are generally higher in men [10]. A large number of factors (e.g., hormones, body fat composition, age, and physical activity) contribute to the aforementioned sex differences in blood lipids and FA; however, emerging evidence suggests that genetic factors may contribute to these sex differences.

Past genome-wide association studies have identified relationships between single nucleotide polymorphisms (SNP) and lipids/FA. These studies suggest that genetics may explain a significant proportion of the variation in circulating lipids and FA. Indeed, 95 SNPs were found to explain up to 14% of the variability observed in blood lipids and, more importantly, these variants correlated with CVD and metabolic risk [11, 12]. Similarly, large-scale genetic investigations reported that SNPs can account for 8–14% of the variance seen in abundant FA in red blood cells (RBC) [13]. While these past studies accounted for sex as a covariate in their analyses, evidence exists to suggest that genomic variation may underlie sex differences in various lipid and FA traits [14, 15]. Consequently, further examination of the relationship between genetic variants and lipid/FA parameters in a sex-specific manner may reveal helpful clues to better understand differences in CVD risk between men and women. Therefore, the goal of the current study was to examine potential sex-genotype interactions between a panel of common candidate SNPs previously associated with blood lipids and/or FA and the lipid/FA profile in two population-based cohorts.

Materials and Methods

Cohort Characteristics

The GONE FISHIN’ cohort comprised 37 men and 57 women between 19–30 years of age, recruited at the University of Guelph between 2012 and 2016 for various omega-3 FA supplementation studies. The present study used only baseline anthropometric, bioclinical, and RBC FA data from these individuals. All participants were deemed healthy and did not report a diagnosed medical condition (e.g., cancer or diabetes mellitus). Individuals using lipid-altering medications were excluded. 35 of 57 females reported using oral contraceptives. These studies were approved by the Human Research Ethics Board at the University of Guelph, and written consent was collected for all participants.

The Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) study is a part of the NHLBI Family Heart Study. GOLDN recruited participants (n = 1327) from families of European descent at two field centers: Minneapolis, MN and Salt Lake City, UT. The objective of the intervention study was to identify genetic factors that determine lipid responses to a high-fat meal challenge and 3-week treatment of fenofibrate (160 mg/per day) [16]. Prior to study visit, participants were requested not to use lipid-lowering medications for at least 4 weeks, fast for at least 8 h, and consume no alcohol for at least 24 h. The study protocol was approved by the Institutional Review Boards at Tufts University, the University of Minnesota, the University of Utah, and the University of Alabama at Birmingham. All participants provided written consent for the study. The present study consisted of a subset of age-matched men (n = 57) and women (n = 61; i.e., age ≤ 30 years old) at baseline from the GOLDN study.

Clinical and Anthropometric Measurements

GONE FISHIN’

Blood collection occurred at the Human Nutraceutical Research Unit at the University of Guelph (Guelph, Ontario). Blood samples were collected from participants via venous puncture after an overnight fast. The blood collection study visit was scheduled during the follicular phase of the women’s menstrual cycle, as lipid measurements were previously reported to be more consistent during this time [17]. Fasted serum samples from each participant were sent to Lifelabs Medical Laboratory Services (Kitchener, Ontario, Canada) for the analysis of TAG, HDL-c, LDL-c, and TC. Anthropometric measures (age, height, body weight, and calculated BMI) were also collected on the day of the visit using standard protocols.

GOLDN

Blood samples were collected from participants after overnight fasting [18]. The lipid profile (e.g., TAG, HDL-c, LDL-c, and TC) was measured using a traditional (enzymatic) method [19]. Height and weight were measured while participants were dressed in an examining gown and wearing no shoes. BMI was calculated as weight in kilograms divided by the square of height in meters [20].

Fatty Acid Analysis of Red Blood Cells

GONE FISHIN’

RBC were separated from plasma following centrifugation at 3000 rpm for 15 min at 4 °C. RBC were then aliquoted into 1.5-mL Eppendorf tubes and stored at −80 °C until analyzed. RBC FA were extracted with a mixture of chloroform and methanol (2:1, v/v) using the methodology established by Folch et al. [21]. Gas chromatography was performed as previously described [22]. FA methyl esters were separated by gas chromatography using an Agilent 6890B gas chromatograph (Agilent Technologies, CA, USA). FA peaks were identified by comparison to retention times of FA methyl ester standards. Individual FA are indicated as a relative percentage (%) of total FA. FA contributing to < 0.25% of the total FA profile were excluded for the present analysis. Enzyme activities were estimated using established product-to-precursor ratios for stearoyl-CoA desaturase (SCD-16 and SCD-18), elongase 6 (ELOVL6), and FA desaturases 1 and 2 (FADS1 and FADS2), as previously reported [23].

GOLDN

RBC FA were extracted with a mixture of chloroform and methanol (2:1, v/v), collected in heptane, and then injected onto a Varian CP7420 100-m capillary column with a Hewlett-Packard 5890 gas chromatograph equipped with an HP6890A autosampler. To separate FA from 12:0 through 24:1n-9, the initial temperature of 190 °C was increased to 240 °C over 50 min [24]. Enzyme activities were estimated as described above.

Selection of Genetic Variants

The panel of candidate SNPs used for the present analysis were selected following an extensive review of the literature for gene variants associated with lipid and/or FA metabolism (Table 2). SNPs selected for the present investigation met several criteria, including: (1) associations were identified in large study populations, (2) used participants primarily of European/Caucasian descent, (3) showed a statistically significant association with a lipid and/or FA parameter, and (4) had been replicated in two or more independent studies. Finally, candidate SNPs had to have a minor allele frequency (MAF) of > 5% in the European population in the 1000 Genomes database (release 17 - Nov 2015).

Table 2.

Red blood cell fatty acid values in men and women in the GONE FISHIN’ and GOLDN cohorts

Common name	Fatty acid	GONE FISHIN’				GOLDN
		Total (n = 94)	Female (n = 57)	Male (n = 37)	p value	Total (n = 118)	Female (n = 61)	Male (n = 57)	p value
Saturated
Myristic acid	14:0	0.54 ± 0.02	0.62 ± 0.02	0.44 ± 0.03	7.22E-5	0.29 ± 0.01	0.31 ± 0.02	0.26 ± 0.01	0.04
Palmitic acid	16:0	22.30 ± 0.12	22.28 ± 0.18	22.33 ± 0.16	0.73	22.62 ± 0.11	22.81 ± 0.16	22.42 ± 0.13	0.15
Stearic acid	18:0	11.95 ± 0.15	11.57 ± 0.15	12.53 ± 0.21	1.61E-3	9.20 ± 0.06	9.12 ± 0.10	9.30 ± 0.07	0.18
Arachidic acid	20:0	0.47 ± 0.02	0.42 ± 0.01	0.55 ± 0.05	0.02	0.20 ± 0.01	0.21 ± 0.01	0.20 ± 0.01	0.40
Behenic acid	22:0	1.31 ± 0.02	1.27 ± 0.02	1.37 ± 0.02	1.89E-3	0.49 ± 0.01	0.51 ± 0.02	0.48 ± 0.02	0.20
Lignoceric acid	24:0	3.91 ± 0.22	3.98 ± 0.36	3.79 ± 0.07	0.05	ND	ND	ND	ND
Monounsaturated
Palmitoleic acid	16:1n-7	0.64 ± 0.03	0.75 ± 0.02	0.44 ± 0.03	5.66E-10	0.36 ± 0.01	0.40 ± 0.01	0.32 ± 0.01	9.97E-4
Oleic acid	18:1n-9	13.40 ± 0.18	13.45 ± 0.12	13.33 ± 0.41	0.30	13.91 ± 0.08	13.82 ± 0.11	13.99 ± 0.13	0.26
Vaccenic acid	18:1n-7	1.27 ± 0.04	1.39 ± 0.10	1.10 ± 0.02	0.18	ND	ND	ND	ND
Erucic acid	22:1n-9	0.59 ± 0.02	0.64 ± 0.02	0.52 ± 0.02	1.96E-4	ND	ND	ND	ND
Nervonic acid	24:1n-9	3.94 ± 0.06	4.01 ± 0.07	3.83 ± 0.13	0.33	1.13 ± 0.03	1.15 ± 0.04	1.11 ± 0.05	0.22
Polyunsaturated
Linoleic acid	18:2n-6	13.79 ± 0.16	13.89 ± 0.19	13.64 ± 0.29	0.40	13.30 ± 0.12	13.14 ± 0.17	13.48 ± 0.17	0.18
α-Linolenic acid	18:3n-3	0.45 ± 0.12	0.47 ± 0.01	0.44 ± 0.10	0.01	0.14 ± 0.01	0.15 ± 0.01	0.13 ± 0.01	0.60
dihomo-γ-linoleic acid	20:3n-6	1.62 ± 0.04	1.59 ± 0.06	1.65 ± 0.08	0.83	1.76 ± 0.03	1.76 ± 0.03	1.76 ± 0.05	0.38
Arachidonic acid	20:4n-6	12.52 ± 0.12	12.24 ± 0.13	12.96 ± 0.21	2.49E-3	13.69 ± 0.08	13.67 ± 0.11	13.71 ± 0.13	0.59
Eicosapentaenoic acid	20:5n-3	0.57 ± 0.02	0.53 ± 0.02	0.64 ± 0.03	1.85E-3	0.42 ± 0.01	0.39 ± 0.01	0.46 ± 0.02	0.01
Adrenic acid	22:4n-6	2.49 ± 0.09	2.65 ± 0.20	2.24 ± 0.05	0.69	ND	ND	ND	ND
Docosapentaenoic acid	22:5n-3	2.01 ± 0.03	1.98 ± 0.04	2.06 ± 0.04	0.33	2.07 ± 0.03	1.99 ± 0.04	2.16 ± 0.03	7.76E-3
Docosahexaenoic acid	22:6n-3	3.33 ± 0.07	3.49 ± 0.09	3.08 ± 0.10	3.04E-3	2.69 ± 0.06	2.88 ± 0.07	2.50 ± 0.08	4.22E-4
Estimates of enzyme activity
SCD-16	16:1n-7/16:0	0.03 ± 0.01	0.03 ± 0.01	0.02 ± 0.01	5.47E-11	0.02 ± 0.01	0.02 ± 0.01	0.01 ± 0.01	1.81E-3
SCD-18	18:1n-9/18:0	1.12 ± 0.08	1.17 ± 0.12	1.06 ± 0.03	0.02	1.52 ± 0.02	1.53 ± 0.02	1.51 ± 0.02	0.60
ELOVL6	18:0/16:0	0.54 ± 0.01	0.52 ± 0.01	0.56 ± 0.01	7.62E-3	0.41 ± 0.01	0.40 ± 0.01	0.42 ± 0.01	0.05
FADS1	20:4n-6/20:3n-6	7.73 ± 0.63	7.70 ± 0.28	7.85 ± 1.55	0.57	8.08 ± 0.15	7.99 ± 0.19	8.19 ± 0.25	0.49
n-3 FADS pathway activity	20:5n-3/18:3n-3	1.28 ± 0.10	1.15 ± 0.10	1.45 ± 0.18	1.07E-3	3.09 ± 0.10	2.67 ± 0.10	3.53 ± 0.17	1.90E-05
n-6 FADS pathway activity	20:4n-6/18:2n-6	0.91 ± 0.02	0.88 ± 0.02	0.95 ± 0.03	0.01	1.04 ± 0.01	1.05 ± 0.02	1.03 ± 0.02	0.51

Fatty acids in women and men are reported as relative % of total fatty acids in RBC. A Mann–Whitney U test was used to determine sex differences (p ≤ 0.05). Data is reported as mean % ± SEM

Bolded values indicate significance in both cohorts following Bonferroni correction

ND not determined

DNA Extraction and Genotyping

GONE FISHIN’

DNA was extracted from saliva for female participants using the Oragene DNA collection kit (DNA Genotek Ontario, Canada), and from whole blood for male participants using the Qiagen Paxgene Blood DNA kit (Qiagen, Toronto, Canada), according to the manufacturer’s instructions. Genotyping was performed using the Sequenom MassArray platform at the Sick Kids Genetic Analysis Facility (Toronto, Canada). A total of 94 DNA samples were analyzed, of which no Mendelian errors were detected. Two DNA samples were randomly selected for replication and 100% concordance was achieved.

GOLDN

In this study, data from GOLDN was used to replicate findings from the GONE FISHIN’ population. The process of genome-wide genotyping in GOLDN has been described in detail [25, 26]. Briefly, the hybrid genotype data of 10 SNPs was used, among which two SNPs (rs174570 and rs688) were genotyped using the Affymetrix Genome-wide 6.0 Array (Affymetrix, Santa Clara, CA, USA). The remaining eight SNPs (rs174537, rs1799883, rs1800775, rs2060792, rs2151916, rs3211956, rs6824447, and rs953413, with the imputation quality score R square > 0.63) were imputed using MaCH software (Version 1.0.16) with the human genome build 36 as reference, and genotyped SNPs that met the following criteria [27]: call rate > 96%, minor allele frequency > 1%, and Hardy–Weinberg equilibrium (HWE) test p > 10⁻⁶.

Statistical Analyses

Prior to data analysis, all bioclinical and anthropometric variables were assessed for normality using the Shapiro–Wilk’s test. A non-parametric Mann–Whitney U test was used to compare differences in anthropometric, clinical, and RBC FA data between men and women. Deviations from HWE were tested for each SNP within the total population, as well as within the male and female subgroups, using an χ² test. This conservative approach ensured that both HWE and MAF were comparable in the total population, as well as within the male and female subgroups. Because of low numbers for homozygote minor alleles for most SNPs, we combined heterozygous and minor homozygous subjects into a single group termed “minor allele carriers”. Linear regression was used to examine the associations between SNPs and lipids and/or FA. Models were adjusted for age and BMI. When appropriate, the genotype-sex interaction was assessed with a two-way analysis of variance (reported as p_int), followed by a post hoc Tukey’s HSD. GraphPad Prism 6 (GraphPad Software, Inc., CA, USA) and JMP 12 Statistical Software (SAS Institute, Cary, NC, USA) were used for all GONE FISHIN’ cohort analyses.

For replication in GOLDN, similar covariates were adjusted in the same genetic model (i.e., heterozygous and minor homozygous subjects were combined into a single group termed “minor allele carriers”). In addition, family relationship was modeled using proc GENMOD in SAS 9.4 (Cary, NC, USA) software.

A p < 0.05 was considered statistically significant. A Bonferroni correction was used to account for multiple testing when investigating differences in RBC FA between men and women. We performed a post hoc power calculation using values for palmitoleic acid and HDL-c, since these were previously reported to differ between men and women. With a sample size of 37 men and 57 women in the GONE FISHIN’ cohort, this study had 99% power to detect differences in FA and blood lipids at a 5% significance level.

Results

Blood Lipids

All blood lipid values were in the normal healthy range in participants from both cohorts, in accordance with American Heart Association guidelines [28]. Compared to the GONE FISHIN’ cohort, values for LDL-c, the TC/HDL-c ratio, and TAG were slightly higher in the GOLDN cohort, while TC and HDL-c levels were slightly lower. As reported in Table 1, women in both cohorts had higher HDL-c levels (p < 0.01), while men had a higher TC/HDL-c ratio (p < 0.01). No other differences in blood lipids were observed between men and women.

Table 1.

Blood lipid measures in the GONE FISHIN’ and GOLDN cohorts

Parameter	GONE FISHIN’				GOLDN
	Total (n = 94)	Female (n = 57)	Male (n = 37)	p value	Total (n = 118)	Female (n = 61)	Male (n = 57)	p value
Age (years)	21.99 ± 0.20	21.63 ± 0.20	22.54 ± 0.37	0.08	22.5 ± 0.45	22.71 ± 0.51	22.28 ± 0.76	0.81
BMI (kg/m2)	24.07 ± 0.35	23.72 ± 0.46	24.61 ± 0.53	0.10	25.64 ± 0.53	26.00 ± 0.82	25.26 ± 0.67	0.78
Total cholesterol (TC; mmol/L)	4.30 ± 0.10	4.39 ± 0.14	4.19 ± 0.13	0.17	3.97 ± 0.07	3.97 ± 0.07	3.97 ± 0.07	1.00
LDL-c (mmol/L)	2.32 ± 0.08	2.23 ± 0.10	2.46 ± 0.12	0.13	2.48 ± 0.07	2.42 ± 0.09	2.55 ± 0.09	0.26
HDL-c (mmol/L)	1.60 ± 0.04	1.76 ± 0.05	1.35 ± 0.05	2.49E-7	1.14 ± 0.03	1.22 ± 0.04	1.06 ± 0.03	5.80E-3
TC/HDL-c ratio	2.86 ± 0.08	2.63 ± 0.09	3.21 ± 0.13	2.51E-4	3.68 ± 0.11	3.39 ± 0.12	3.99 ± 0.19	0.02
TAG (mmol/L)	0.96 ± 0.04	1.03 ± 0.06	0.84 ± 0.04	0.08	1.11 ± 0.07	1.04 ± 0.09	1.17 ± 0.10	0.36

Data is reported as mean ± SEM

Bolded values indicate significant sex differences within each cohort (p < 0.05)

BMI body mass index, HDL-c high-density lipoprotein cholesterol, LDL-c low-density lipoprotein cholesterol, TAG triacylglycerol

RBC Fatty Acids

Palmitic, stearic, oleic, linoleic, and arachidonic acid were the dominant FA in RBC in both cohorts (Table 2). Measurements for four FA were not available in the GOLDN cohort (i.e., lignoceric, vaccenic, erucic, and adrenic acid); therefore, these FA were not considered further.

To identify sex differences in RBC FA, we performed three comparisons that varied in their degree of stringency. First, we investigated which RBC FA differed significantly between men and women in both cohorts after accounting for multiple testing (i.e., high stringency). Only palmitoleic acid (16:1n-7) met this criterion, with women having higher levels (Table 2). Second, we next considered FA that were significantly different between men and women, but for which this difference met our significance threshold corrected for multiple testing in only one of the two cohorts (i.e., medium stringency). As shown in Table 2, myristic acid (14:0) and DHA (22:6n-3) were higher in RBC in women, while eicosapentaenoic acid (EPA, 20:5n-3) levels were lower in women. Finally, we also considered those FA that were found to be statistically different between sexes in only one of the two cohorts in the absence of a multiple testing correction (i.e., low stringency). SFA (stearic acid—18:0; behenic acid—22:0) were present at higher levels in men, while PUFA (ALA—18:3n-3; docosapentaenoic acid—22:5n-3) were lower (Table 2). No other differences in FA were observed between men and women.

Desaturation and Elongation Activities

Estimated enzyme activities were generally comparable between the two cohorts. The only exception was with the estimate for “n-3 FADS pathway activity”, which was noticeably higher in GOLDN participants, driven by their lower levels of ALA. Similar trends were seen for the “n-6 FADS pathway activity” estimate.

Using a similar approach with regards to analysis stringency (as outlined above for individual FA), we first examined estimated enzyme activities that were statistically significant after accounting for multiple testing in both cohorts. The SCD desaturation index (SCD-16) was higher in women, while “n-3 FADS pathway activity” was higher in men (Table 2). The ELOVL6 estimate was marginally higher in men, but met our criteria for multiple testing in only one of the two cohorts. The remaining estimates (SCD-18, FADS1 and n-6 FADS pathway activity) were not significantly different between men and women in either cohort.

Characteristics of Candidate SNPs

To investigate genotype-sex interactions, we selected 23 candidate SNPs in 17 genes that were previously reported in the literature to be associated with lipid and/or FA metabolism (Table 3). Specifically, we examined the relationship between SNPs in FASN, SREBP1C, APOA5, CD36, HSL, LPL, CETP, and PPARα with blood lipids, and the relationship between SNPs in FADS1, FADS2, ELOVL2, ELOVL6, SCD1, FASN, FABP2, GPR120, and CD36 with RBC FA. MAFs for all SNPs ranged between 6 and 50% (Table 3). Four SNPs were not in HWE (rs688, rs2297508, rs61866610, 10883463) and therefore not considered further.

Table 3.

Candidate SNP panel determined using the EUR population in the 1000 Genomes database, release 17 - Nov 2015

SNP	Gene	MAF 1000 Genomes	GONE FISHIN’			GOLDN			Investigated SNP associations with lipid and/or FA
			MAF total	MAF males	MAF females	MAF total	MAF females	MAF males
rs174537 (G/T)	FADS1	0.35	0.29	0.32	0.27	0.38	0.34	0.42	Fatty acids
rs174575 (C/G)	FADS2	0.26	0.28	0.35	0.24	NA	NA	NA	Fatty acids
rs174570 (C/T)	FADS2	0.16	0.15	0.18	0.14	0.16	0.15	0.17	Fatty acids
rs953413 (A/G)	ELOVL2	0.44	0.50	0.55	0.46	0.40	0.39	0.39	Fatty acids
rs9393903 (G/A)	ELOVL2	0.24	0.27	0.28	0.25	NA	NA	NA	Fatty acids
rs6824447 (A/G)	ELOVL6	0.47	0.48	0.42	0.48	0.47	0.44	0.48	Fatty acids
rs2060792 (T/C)	SCD1	0.32	0.27	0.20	0.32	0.26	0.22	0.31	Fatty acids
rs10883463 (T/C)	SCD1	0.05	0.09	0.19	0.08	NA	NA	NA	Fatty acids
rs2229422 (A/G)	FASN	0.31	0.35	0.27	0.39	NA	NA	NA	Lipids and fatty acids
rs2297508 (A/G)	SREBP-1C	0.41	0.45	0.46	0.44	NA	NA	NA	Lipids
rs662799 (A/G)	APOA-5	0.08	0.07	0.08	0.06	NA	NA	NA	Lipids
rs1799883 (C/T)	FABP-2	0.27	0.30	0.26	0.33	0.25	0.21	0.28	Fatty acids
rs61866610 (C/T)	GPR120	0.04	0.08	0.07	0.09	NA	NA	NA	Fatty acids
rs3211956 (T/G)	CD36	0.08	0.06	0.05	0.06	0.09	0.11	0.07	Lipids and fatty acids
rs2151916 (T/C)	CD36	0.37	0.35	0.41	0.31	0.44	0.42	0.46	Lipids and fatty acids
rs1206034 (G/A)	HSL	0.35	0.36	0.32	0.38	NA	NA	NA	Lipids
rs320 (T/G)	LPL	0.29	0.31	0.34	0.30	NA	NA	NA	Lipids
rs328 (C/G)	LPL	0.13	0.14	0.14	0.14	NA	NA	NA	Lipids
rs1800775 (A/C)	CETP	0.49	0.49	0.47	0.49	0.43	0.47	0.40	Lipids
rs688 (C/T)	LDLR	0.44	0.38	0.36	0.39	0.46	0.45	0.47	Lipids
rs1800206 (C/G)	PPAR-α	0.05	0.08	0.08	0.08	NA	NA	NA	Lipids and fatty acids
rs708272 (G/A)	CETP	0.43	0.41	0.43	0.39	NA	NA	NA	Lipids
rs5882 (A/G)	CETP	0.33	0.37	0.34	0.39	NA	NA	NA	Lipids

SNP single nucleotide polymorphism, MAF minor allele frequency, NA not available, FA fatty acids

Gene-Lipid Associations

We first investigated associations between SNPs and blood lipids in men and women combined in each cohort. Statistically significant associations between rs1800775 in CETP with HDL-c levels (p = 2.49E-7) and the TC/HDL-c ratio (p = 2.51E-4) were observed in the GONE FISHIN’ cohort. Specifically, we found that major allele carriers (AA) had significantly higher HDL-c levels when compared to their minor allele carrier counterparts (CA+CC), while the opposite relationship was seen for TC/HDL-c. However, these associations failed to replicate in GOLDN (p = 0.29 and p = 0.97, respectfully). In FASN, rs2229422 was significantly associated with TAG levels in the GONE FISHIN’ cohort, with major allele carriers (AA) having lower TAG compared to minor allele GA/GG carriers (p = 0.04). However, we were unable to confirm this finding, as this SNP was not genotyped in the GOLDN cohort. None of the other candidate SNPs were associated with any of the examined lipid traits.

Genotype x Sex Interactions Related to Blood Lipids

Overall, no consistent genotype-sex interactions were detected with blood lipids in the present study. As mentioned above, an association was identified between rs1800775 in CETP and the TC/HDL-c ratio in the GONE FISHIN’ cohort. We also found a significant genotype-sex interaction for this association in the same cohort (p_int = 0.03). Specifically, male homozygotes for the major allele (AA) had a significantly lower TC/HDL-c ratio compared to male carriers of the minor allele (CA/CC). No difference between genotypes was observed between women. However, this genotype-sex interaction was not replicated in the GOLDN cohort.

Gene-Fatty Acid Associations

We next investigated associations between SNPs and RBC FA in the combined male and female population of both cohorts. In FADS1, rs174537 was significantly associated with arachidonic acid levels in both cohorts (p < 0.01), with major allele carriers (GG) having higher levels compared to minor allele carriers (GT+TT; Fig. 1a). We also observed that this same SNP was significantly associated with dihomo-γ-linoleic acid (DGLA, 20:3n-6) in both cohorts (p < 0.01), with major allele carriers having lower levels compared to minor allele carriers (Fig. 1b).

Fig. 1.

Significant associations between FA and SNPs in FADS1 and CD36. Relative percentage for AA (20:4n-6) (a) and DGLA (20:3n-6) (b) in participants stratified as major (GG) or minor (GT+TT) carriers of rs174537 in FADS1. (c) FADS1 activity estimate (i.e., 20:4n-6/20:3n-6) in participants stratified as major (TT) or minor (GT+GG) for rs3211956 in CD36. Data is reported as mean% ± SEM. p values correspond to post hoc analyses following a two-way ANOVA

Finally, major allele carriers (TT) for the rs3211956 SNP in CD36 had a higher estimated FADS1 ratio (i.e., AA/DGLA) compared to minor allele carriers (GT+GG) in both cohorts (Fig. 1c).

Genotype x Sex Interactions Related to Fatty Acids

No genotype-sex interactions were detected between any of the candidate SNPs investigated in the current study and RBC FA levels.

Discussion

The present study investigated if sex differences in blood lipids and FA were associated with common genetic variants previously reported in the literature. The primary findings from the current study were: (1) HDL-c was higher and the TC/HDL-c ratio was lower in women compared to men, (2) palmitoleic acid and the SCD desaturation index were higher in RBC in women compared to men, and (3) despite finding weak associations in the total population between SNPs in CETP, FASN, and FADS1 with HDL-c, TAG, and specific FA levels, respectively, we found no evidence that these traits were modulated by the candidate SNPs in a sex-specific manner. Overall, our findings suggest that the SNPs investigated in the current study do not contribute significantly to sex differences in blood lipids and FA in young adults.

Sex Differences in Blood Lipids and Fatty Acids

Risk of CVD mortality is two to five times greater in men versus women [29]. In general, risk factors for CVD, which include HDL-c and TC, are more favorable in women compared to men. In agreement with previous reports [29–31], we found that women had higher HDL-c levels and a lower TC/HDL-c ratio compared to men. These differences align with previous findings in the Framingham Offspring Study [32]. Interestingly, a past study by Freedman et al. found that not only was HDL-c higher in women than men, but so was the proportion of larger anti-atherosclerotic HDL particles [32]. Together, this would confer a reduced risk of CVD in women. Further, these aforementioned sex differences were most apparent in younger adults compared to older adults. In contrast to Freedman et al. [32], we did not observe sex differences in LDL-c and TAG levels. It is possible that these discrepant findings stem from differences in sample size and demographics (i.e., Freedman et al. had a larger sample and investigated associations in older adults compared to the current study).

We also observed sex differences in several RBC FA and estimates of enzyme activity. Most notably, we found that women had higher levels of RBC palmitoleic acid. While palmitoleic acid can be obtained in low quantities through the consumption of plant oils and animal fats, it is primarily produced de novo from the delta-9 desaturation of palmitic acid [33]. In agreement with increased palmitoleic acid levels, we also found that women had a higher SCD-16 desaturation index. Together, this suggests that females have increased SCD activity compared to men. Stark et al. previously reported that post-menopausal women had lower relative levels of serum palmitoleic acid compared to pre-menopausal and post-menopausal women receiving hormone therapy [34]. This suggests that the higher levels of palmitoleic acid and the SCD-16 desaturation index in young women may be related to hormone differences between the two sexes. Additionally, blood levels of palmitoleic acid have been proposed to serve as a marker of adipose tissue lipolysis and hepatic de novo synthesis [35]. Consequently, RBC palmitoleic acid levels may serve as a potential blood indicator for a person’s overall metabolic health.

We also found sex differences in EPA and DHA levels, with women having less EPA and more DHA in RBC. Interestingly, when estimating the activity of the FADS pathway using the product-to-precursor ratio of EPA/ALA, we observed that men had a higher FADS pathway activity. While seemingly contradictory, these findings can be reconciled when considering past studies that investigated ALA conversion into DHA using tracers. Specifically, it was shown that women have a higher capacity to convert ALA into DHA [36]. Consequently, this means that women would be expected to have higher levels of DHA in relation to EPA. In contrast, men have a lower ability to synthesize DHA, which would lead to an accumulation of EPA. This differential pathway activity is particularly intriguing given emerging evidence showing that EPA and DHA have distinct effects on blood lipids, glycemic control, and inflammation [37, 38].

Finally, we also found that women had higher levels of RBC myristic acid compared to men. This is noteworthy given that previous studies have reported positive correlations between circulating myristic acid levels and CVD risk [16, 39, 40]; however, these studies did not stratify their cohorts by sex when investigating these associations. As such, caution is warranted when interpreting our finding that women had higher RBC myristic acid levels. Indeed, the relative amounts of myristic acid compared to other saturated fats (e.g., palmitic and stearic acids) are considerably lower in RBC, and no sex differences were observed with these more abundant SFA. Moreover, when considered collectively, young females in our study had a more favourable CVD risk profile compared to men. Nevertheless, future studies examining sex differences in FA profiles should consider monitoring diet intake and de novo lipogenesis, both of which could explain these minor sex differences in myristic acid.

The Relationship Between Blood Lipids, Genotype, and Sex Differences

We examined a panel of SNPs located in eight genes known to influence blood lipids, and found that only rs1800775 in CETP was significantly associated with HDL-c levels and the TC/HDL-c ratio. While rs1800775 was genotyped in both the GONE FISHIN’ and GOLDN, the identified associations were not consistent between the two cohorts. The lack of reproducibility could stem from a multitude of reasons that include differences in lifestyle and environmental factors, small sample sizes, and demographics between the two cohorts. Nevertheless, our finding in the GONE FISHIN’ cohort aligns with previous reports by other research groups and thus warrants further discussion.

The CETP gene encodes the cholesterol ester transfer protein, which plays a critical role in reverse cholesterol transport [2]. Past research has shown that homozygote carriers of the major allele (AA) for the rs1800775 SNP in CETP have significantly higher HDL-c levels compared to their minor allele counterparts (CA+CC) [41, 42]. An inverse relationship was also seen between this SNP and the TC/HDL-c ratio [43]. In our study, we found similar associations in GONE FISHIN’, but not GOLDN. While rs1800775 in CETP might be valuable for CVD risk prediction in the general population, we did not find evidence that this SNP had a differential effect on HDL-c levels in men and women. Nevertheless, past evidence suggests a relationship between SNPs in the CETP gene and sex differences in blood lipids may exist. For example, Papp et al. revealed that a haplotype consisting of the rs5883 and rs9930761 SNPs in CETP was associated with HDL-c levels and CVD risk in men, but not women [44]. Moreover, Anagnostopoulou et al. also reported an interaction between sex and two other SNPs in CETP, rs708272 (also known as TaqIB), and rs5882 (also known as I405V), on postprandial TAG response following an oral fat tolerance test [45]. Collectively, these previous reports highlight that further investigation is necessary to clarify the sex-specific effects of genetic variation in CETP on both fasting and postprandial lipids.

The Relationship Between RBC Fatty Acids, Genotype and Sex Differences

There is considerable evidence in the literature that SNPs in the FADS1 gene are associated with serum, plasma, RBC, and tissue FA profiles [46]. The present study replicated previous findings between the rs174537 SNP in FADS1 and specific FA, but found no evidence that this association varied in a sex-specific manner. FADS1 encodes the delta-5 desaturase (D5D), which regulates the conversion of DGLA into AA [46]. Previous reports have shown that individuals carrying minor alleles in FADS1 SNPs have lower desaturase activity compared to major allele carriers [46]. As expected, minor allele carriers (GT+TT) for the rs174537 SNP in FADS1 had lower RBC AA levels compared to major allele carriers (GG), and correspondingly more DGLA. This observation may have implications for CVD risk, because this same SNP was previously shown to not only influence AA synthesis, but also pro-inflammatory oxylipin production in whole blood [47]. However, we did not find any evidence of a sex-genotype interaction between FADS1 and omega-6 PUFA in RBC.

Interestingly, we also observed a higher FADS1 activity estimate (AA/DGLA) in minor allele carriers (TT) of the rs3211956 SNP in CD36 compared to major allele carriers (GT+GG). The CD36 gene encodes an integral membrane protein, which binds and facilitates the uptake of FA from blood into tissues around the body [48]. Although not possible to definitively conclude from our data, this association may suggest a degree of FA specificity by CD36. In other words, CD36 protein in minor allele carriers for the rs3211956 SNP may have a preferential affinity for DGLA uptake into tissues compared to AA, which would consequently lead to an increase in the FADS1 activity estimate in RBC. Although we did not detect a sex-genotype interaction with the rs3211956 SNP (nor has one been reported previously in the literature to the best of our knowledge), sex differences in hepatic CD36 gene expression have been demonstrated by Stahlberg et al. [49]. It may be valuable to continue investigating the role of CD36, both at the level of SNPs as well as gene/protein expression, as this may uncover differences in FA transport mechanisms between men and women that could alter CVD risk.

Strengths and Limitations

The present study has several strengths and limitations that should be considered. A major strength of this study was the use of two independent age-matched cohorts to investigate sex differences in lipids and FA in young adults. However, we acknowledge that the sample sizes of the age-matched cohorts were small, and that the sex distribution was not equivalent. Future studies with a comparable number of men and women, as well as increased sample size, would improve the power to identify sex-genotype interactions. From a technical perspective, the two studies used different approaches for genotyping. As such, not all of the candidate SNPs investigated in the GONE FISHIN’ cohort were available in the GOLDN cohort. Finally, we acknowledge that investigating the influence of these common SNPs in diseased or at-risk populations may reveal sex differences that are not apparent in healthy young adults.

In conclusion, findings from the current study have both confirmed and identified sex differences in specific blood lipids/FA in RBC between young, healthy male and female adults. Our results do not support the hypothesis that these differences are related to SNPs in key candidate genes associated with lipid and FA metabolism. However, previous studies in the literature suggest that such sex-genotype interactions might exist. Further investigations examining factors that contribute to sex differences in blood lipids and FA will continue to advance our understanding of differences in CVD risk between men and women.

Abbreviations

CD36

Cluster of differentiation 36

CETP

Cholesteryl ester transfer protein

CVD

Cardiovascular disease

FA

F atty acids

FADS

Fatty acid desaturase

GOLDN

The Genetics of Lipid Lowering Drugs and Diet Network

HDL-c

High-density lipoprotein cholesterol

HWE

Hardy–Weinberg equilibrium

LDL-c

L ow-density lipoprotein cholesterol

MAF

Minor allele frequency

MUFA

Monounsaturated fatty acids

PUFA

Polyunsaturated fatty acids

RBC

Red blood cell

SCD

Stearoyl-CoA desaturase

SFA

Saturated fatty acids

SNP

Single nucleotide polymorphism

TAG

Triacylglycerol

TC

T otal cholesterol