Clinica Chimica Acta 431 (2014) 244–249

Contents lists available at ScienceDirect

Clinica Chimica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /c l inch im

Gender-specific associations of the APOA1−75GNA polymorphismwithseveral metabolic syndrome components in Turkish adults

Neslihan Coban a,1, Altan Onat b,1, Filiz Guclu-Geyik a,1, Evrim Komurcu-Bayrak a,1,Gunay Can c,1, Nihan Erginel-Unaltuna a,⁎,1

a Department of Genetics, Institute for Experimental Medicine, Istanbul University, Istanbul, Turkeyb Department of Cardiology, Cerrahpasa Medical Faculty, Istanbul University, Istanbul, Turkeyc Cerrahpaşa Medical Faculty, Depatrment of Public Health, Istanbul University, Istanbul, Turkey

⁎ Corresponding author at: Department of GeneticMedicine, Istanbul University, Vakif Gureba Cad. 3408Tel.: +90 212 4142200 33324; fax: +90 212 5324171.

E-mail addresses: neslihancoban@yahoo.com (N. Coba(A. Onat), filiz.geyik@gmail.com (F. Guclu-Geyik), evrimb(E. Komurcu-Bayrak), alpincan@yahoo.fr (G. Can), nihane(N. Erginel-Unaltuna).

1 These authors contributed equally to the work.

0009-8981/$ – see front matter © 2014 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.cca.2014.01.017

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 25 September 2013Received in revised form 10 January 2014Accepted 10 January 2014Available online 6 February 2014

Keywords:Apolipoprotein A1PolymorphismHDL-CAtherogenic dyslipidemiaBlood pressure

Background: Variations in the apolipoprotein A-1 (APOA1) gene, a determinant of plasma high-density lipopro-tein cholesterol (HDL-C) and apoA-I levels, may contribute to cardiovascular diseases. We evaluated the effectsof a promoter polymorphism (−75GNA) in the APOA1 gene on metabolic syndrome (MetS) components in aTurkish population sample.Methods: Randomly selected 1515 Turkish adults (age 49.9 ± 11.8 years, 785 females) were genotypedfor −75GNA polymorphism using hybridization probes in Real-Time PCR LC480 device. MetS and atherogenicdyslipidemia were defined using the criteria of ATP III.Results: The −75AA genotype prevailed in 3.9% of men and 2.4% of women, and was independently associatedwith significantly higher HDL-C concentrations. Independent associations with the −75GA genotype existedonly in men: higher diastolic and systolic blood pressure (BP) levels (p b 0.05) were observed in male −75GAheterozygotes. Logistic regression revealed that the GA genotype confers elevated risk for atherogenic dyslipid-

emia (OR = 1.57, 95% Cl 1.06–2.3) after adjustment for associated risk factors. Independent associations withatherogenic dyslipidemia or elevated BP did not emerge in women.Conclusion: APOA1 −75GNA polymorphism is independently related to HDL-C concentrations. Independentassociations of the −75GA genotype with elevated BP and atherogenic dyslipidemia were confined to men.These gender-modulated associations suggest novel gene–gender–environmental interactions.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Themetabolic syndrome (MetS) is characterized by the clustering ofseveral factors: abdominal obesity, insulin resistance and/or glucose in-tolerance, atherogenic dyslipidemia, increased systolic and diastolicblood pressure and a pro-inflammatory/pro-thrombotic state [1].Atherogenic dyslipidemia related to MetS is a common metabolicdisorder consisting of elevated plasma triglyceride (TG) and, decreasedconcentrations of high-density lipoprotein cholesterol (HDL-C) is asso-ciated with increased risk of coronary heart disease (CHD). The molar

s, Institute for Experimental0 Sehremini, Istanbul, Turkey.

n), alt_onat@yahoo.com.trayrak@yahoo.comrginel@yahoo.com

ghts reserved.

ratio of fasting TG/HDL-Cwas shown to be a strong independent predic-tor of myocardial infarction [2], and a nominal ratio in excess of 3.6 hasbeen reported to be strongly associated with MetS [3].

Apolipoprotein (apo)A-I is the major surface protein constituent ofHDL-C and plays a vital role in lipid transport and metabolism. Geneticstudies have identified polymorphisms and mutations in the severalgenes associated with HDL-C and apoA-I levels [4]. Genome-wide asso-ciation studies (GWAS) of HDL-C concentrations have previously iden-tified several loci with genome-wide significant associations nearcandidate genes [5]. One of these is the APOA1 locus that maps to thelong arm of chromosome 11 [4]. Overexpression of the human APOA1gene in mice increased plasma HDL-C concentrations and protectedthe animals from the development of diet-induced atherosclerosis [6].

The APOA1 gene polymorphisms have been found to be strongly as-sociated with variation in MetS traits such as serum apoA-I, HDL-C andglucose concentrations [7–10]. A common G-to-A transition located 75base pairs upstream from the transcription start site of the APOA1gene has been studied extensively [7]. Moreover, increased transcrip-tion efficiency was observed in −75A allele carriers in comparisonwith G allele homozygotes [11]. Individuals carrying the −75A allele

245N. Coban et al. / Clinica Chimica Acta 431 (2014) 244–249

have been shown to have higher concentrations of apoA-I and HDL-C[8–10,12]; however, the results have been inconsistent and incon-clusive, with a few studies reporting either no association [13,14]or negative association [15] between APOA1 −75GNA polymor-phism and plasma lipids. In addition, APOA1 −75AA genotype wasreported to be associated with increased risk of type 2 diabetes[16]. Nonetheless, −75GG genotype/G allele has been found to bestrongly associated with myocardial infarction [17], hypertension[18] and MetS components [19], but also with decreased bloodpressure levels [20].

We, therefore, examined the relation of the−75GNApolymorphismwith MetS traits in a randomly selected 1515 sample of the TurkishAdult Risk Factor Study (TARF) cohort, representative of Turkish adults[21].

2. Subjects and methods

2.1. Study population

The design andmethodology of the TARF Study have been previous-ly described [21]. It involves a random sample of the Turkish adultpopulation, representatively stratified for sex, age, geographical regionsand for rural–urban distribution. Participants were recruited fromrandomly selected communities using a probability–proportionate-to-size method and they were visited during the 8 follow-up surveysthrough 2010. Data were obtained for history of the past years via aquestionnaire, physical examination of the cardiovascular systemand recording of a resting electrocardiogram. Randomly selected1515 subjects (730 males and 785 females) were examined fortheir APOA1 −75GNA genotypes, based on the availability of determi-nations of fasting concentrations of HDL-C, triglyceride and glucose.Study subjects were unrelated and gave written consent to participatein the study after being informed of its nature. The study protocolwas approved by the Ethics Committee of the Istanbul Medical Faculty,Istanbul University.

2.2. Definitions

Obesity was defined as a BMI ≥30 kg/m2. Individuals with diabe-tes were diagnosed with criteria of the American Diabetes Associa-tion [22], namely when plasma fasting glucose was N126 mg/dL (or2-h postprandial glucose N200 mg/dL) and/or the current use ofdiabetes medication. Oral glucose tolerance test was not performed.Individuals with MetS were identified when 3 out of the 5 criteria ofthe National Cholesterol Education Program (ATP III) [1] were met,modified for prediabetes (fasting glucose 100–125 mg/dL) [23] andfurther for abdominal obesity using as cutpoint ≥95 cm (instead of102 cm) in men, as assessed in the Turkish Adult Risk Factor study[24]. Atherogenic dyslipidemia was defined by high (N150 mg/dL)fasting triglyceridemia and low concentrations (b40/b50 mg/dL) ofHDL-C [1]. Physical activity was graded by the participant himselfinto four categories of increasing order with the aid of a scheme [21].Grades I and II were considered in the present study as low physicalactivity.

2.3. Measurement of risk factors

Current smokers, never and former smokers formed the categoriesin cigarette smoking. Anyone consuming alcohol once a week or morewas considered as an alcohol user. Weight was measured withoutshoes in light indoor clothes using a scale. Body mass index (BMI) wascomputed as weight divided by height squared (kg/m2). Waist circum-ference was measured with a tape (Roche LI95 63B 00), the subjectstanding and wearing only underwear, at the level midway betweenthe lower rib margin and the iliac crest.

Blood samples were collected after an 11-hour or longer fasting.Samples were shipped to Istanbul to be stored at−75 °C, until analyzedat a central laboratory. Serum concentrations of total cholesterol (TC),fasting triglycerides, glucose, and HDL-C (directlywithout precipitation)were determined using enzymatic kits from Roche Diagnostics with aHitachi 902 autoanalyzer. Concentrations of apoA-I and B weremeasured by Behring kits and nephelometry (BN Prospec, BehringDiagnostics, Westwood, MA).

2.4. Genetic analysis

2.4.1. Determination of the APOA1−75GNA genotypesDNA was extracted from peripheral blood leucocytes using a

QIAmp®DNAMaxi KIT (Qiagen, Hilden, Germany). TheDNA concentra-tions have been standardized and stored in an 8 × 12 format at−20 °C.Unselected 1515 subjects (730 male and 785 female) were examinedfor their APOA1 −75GNA genotypes. Genotyping was performed usinghybridization probes in Real-Time PCR LightCycler® 480 device. DNAamplification was set up in 96 well plates (ABGENE Ltd.) Typical 10 μLPCR reaction consisted of 2 μL LightCycler® 480 genotyping masterready mix (Roche), 0.15 μL probes and 0.2 μL primers, 5.3 μL distilledwater. 2.5 ng genomic DNA was added to the PCR mixture. PCR wascarried out on LightCycler® 480 using the following conditions: 95 °Cfor 10min, 95 °C for 5 s, 56 °C for 5 s, and 72 °C for 7 s (45 cycles).Melt-ing curve analysis was assessed using the LightCycler® 480 genotypingsoftware.

2.5. Statistical analysis

The genotypic distributions were compared using the chi-squaretest. Hardy–Weinberg equilibriumwas computed to the expected geno-type distribution. Due to skewed distributions fasting triglyceride waslogarithmically transformed for analyses and expressed as geometricmeans and standard deviation (SD). Multinominal regression analyseswere performed for interactions between SNP and subgroups. Interac-tions between SNP and the studied traits were examined by using uni-variate analysis. Two-tailed t test and analysis of variance (ANOVA)test were used to compare continuous variables and expressed asmeans and SD, while categorical variables were compared using thechi-square test. ANOVA was employed to compare values among geno-types. One-way analysis of covariance (ANCOVA) was performed withthese values as dependent variables and the following factors as inde-pendent variables: APOA1 genotype, age, waist circumference, lipidlowering medication, current smoking and diabetes mellitus in sub-groups (men and women). Logistic regression models were used forAPOA1 genotype to derive maximum odds ratio (OR) estimates andassociated 95% confidence intervals (CIs), adjusted for several environ-mental factors as confounders. A p-value b0.05 was considered signifi-cant, as was a 0.05 level in multiple testing after application of theBonferroni correction. All statistical analyses were performed usingWindows SPSS version 10.0 software.

3. Results

3.1. Study characteristics

The biometric parameters and characteristics of the participants ofthe study population are shown in Table 1, by gender. The samplecomprised of selected 1515 (mean age; 49.9 ± 11.8, 48.2% male)Turkish adults. The prevalence of MetS, CHD, atherogenic dyslipid-emia, obesity, hypertension and type 2 diabetes was around 49.0%,7.3%, 28.9%, 32.8%, 46.0% and 7.9% of the participants, respectively. Ofall the subjects, 3.6% were already being treated with lipid-loweringdrugs.

Table 1Characteristics of the study sample, by gender.

Characteristics Men (n = 730) Women (n = 785) p-value

mean ± SD mean ± SD

Age (years) 50.4 ± 11.9 49.5 ± 11.8 NSBMI (kg/m2) 27.1 ± 4 29.4 ± 5.5 b0.001Waist circumference (cm) 94.6 ± 10.9 91.1 ± 12.2 b0.001Total cholesterol (mg/dL) 186.6 ± 38.3 193.2 ± 39.04 0.001HDL-cholesterol (mg/dL) 38.7 ± 11.5 46.2 ± 12.5 b0.001LDL-cholesterol (mg/dL) 114.3 ± 32.4 119.2 ± 33.6 0.005Fasting triglyceride (mg/dL)⁎ 144.5 × 1.73 120.2 × 1.66 b0.001Fasting glucose (mg/dL) 99.6 ± 31.7 98.2 ± 28.8 NSApolipoprotein A-I (mg/dL) 129.03 ± 26.9 145.5 ± 29.7 b0.001Apolipoprotein B (mg/dL) 115.2 ± 36.02 113.7 ± 32.1 NSApoA-I/HDL-C ratio 3.51 ± 1.04 3.29 ± 0.87 b0.001Systolic blood pressure (mm Hg) 127.9 ± 22.7 133.8 ± 26.6 b0.001Diastolic blood pressure (mm Hg) 81.9 ± 13.5 83.3 ± 14.4 0.047Coronary heart disease, % (n) 8.2 (60) 6.5 (51) NSMetabolic syndrome, % (n) 49.0 (358) 48.9 (384) NSDiabetes mellitus, % (n) 8.6 (63) 7.3 (57) NSHypertension, % (n) 40.5 (296) 51.1 (384) b0.001Obesity, % (n) 22.2 (150) 42.5 (314) b0.001Atherogenic dyslipidemia, % (n) 32.9 (229) 25.2 (192) 0.001Current smokers, % (n) 46.7 (326) 17.6 (134) b0.001Alcohol consumption, % (n) 18.8 (131) 0.7 (5) b0.001Low physical activity, % (n) 53.0 (343) 79.1 (551) b0.001Lipid lowering drug usage, % (n) 2.7 (20) 4.3 (34) NSDrug usage for hypertension, % (n) 12.1 (88) 24.3 (191) b0.001Usage of antidiabetic drugs, % (n) 5.9 (43) 5.9 (46) NSMenopausal status, % (n) – 43.9 (347) –

NS, not significant. ⁎Geometric mean values. p-values were obtained by two-tailed t-testsand the chi-square tests.

246 N. Coban et al. / Clinica Chimica Acta 431 (2014) 244–249

3.2. Allele and genotype distributions of APOA1 −75GNA polymorphism

The genotype frequencies of the APOA1−75GNA polymorphismweredetermined for 1515 participants of the TARF study population. The dis-tribution of the APOA1−75GNA polymorphism in this adult populationwas 71.4% (n = 1081), 25.5% (n = 386) and 3.1% (n = 48) for the GG,GA and AA genotypes, respectively (Table 2). The −75AA genotypeprevailed in 3.9% of men and 2.4% of women. The−75A allele frequencywas found to be 0.16 and the carrier ship (GA + AA) of rare allele wasfound to be 0.28. Genotype distribution was in Hardy–Weinberg equilib-rium for our study population.

3.3. Association of APOA1−75GNA polymorphism with lipids

The relationship between concentrations of biochemical variablesand the −75GNA polymorphism of the APOA1 gene was examinedafter exclusion of the subjects using lipid lowering drugs (n = 54) inANOVA and Bonferroni adjustment for age, waist circumference,lipid lowering medication, current smoking and diabetes mellitus(Table 3). In women, the −75AA genotype was associated with higherserum concentrations of HDL-C (GG: 45.5 ± 0.52, GA: 47.7 ± 0.86,

Table 2Genotype and allele frequencies of APOA1 −75GNA polymorphism in the study sample.

Variables Study population Males Females

Genotype frequencya

GG, % (n) 71.4 (1081) 71.8 (524) 71 (557)GA, % (n) 25.5 (386) 24.3 (177) 26.6 (209)AA, % (n) 3.1 (48) 3.9 (29) 2.4 (19)GA + AA, % (n) 28.6 (434) 28.2 (206) 29 (228)

Allele frequencyG, % 84.1 83.9 84.3A, % 15.9 16.1 15.6

a Frequency of alleles was computed using gene-counting method. The associationbetween the genotype frequencies and gender was compared using the Chi-square test(p N 0.05).

AA: 52.2 ± 2.79 p = 0.008). Borderline associations with higher HDL-C concentrations (p = 0.057) were observed in men. In ANCOVA test,the −75AA genotype was associated with lower apoA1/HDL-C ratio(GG: 3.44 ± 0.04, GA: 3.34 ± 0.06, AA: 2.99 ± 0.17, p = 0.027) inthe total sample and a borderline association (p = 0.064) in men. Wefound no association of APOA1 −75GNA polymorphism with serumapoA1, ApoB, LDL-C, total cholesterol and TG concentrations in thispopulation (data not shown).

3.4. Associations of APOA1 −75GNA polymorphism with serum glucoseconcentration

The relationship between fasting glucose concentrations andthe −75GNA polymorphism of the APOA1 gene in a total of 1426 indi-viduals (after the exclusion of 89 subjects using antidiabetic drugs)are provided in Table 3. InAPOA1−75GA genotype, after Bonferroni ad-justment for age, waist circumference, lipid lowering medication, cur-rent smoking, diabetes mellitus, fasting glucose concentrations werelower inmen (p= 0.021)while statistical significancewas not reachedin women. In relation to glucose concentration, the APOA1 −75GNApolymorphism showed significance in gender-by-genotype interaction(p = 0.017).

3.5. Associations of APOA1 −75GNA polymorphism with serum bloodpressure levels

Systolic blood pressure (SBP) and diastolic blood pressure (DBP)levels, adjusted for age, waist circumference, hypertension medication,current smoking, diabetes mellitus and atherogenic dyslipidemia,differed significantly across genotypes in men but not in women(Table 3). Mean values were higher in men with the −75A allelecarriers than in G homozygotes (DBP 84.0 vs. 81.0 mm Hg, p b 0.01and SBP 129.9 vs. 127.1 mm Hg, p = 0.089).

3.6. Associations of APOA1 −75GNA polymorphism with atherogenicdyslipidemia

In relation to atherogenic dyslipidemia, the APOA1 −75GNApolymorphism had significant gender-by-genotype interaction (p =0.001). As seen in Table 4, this polymorphismwas significantly associat-ed with dyslipidemia in men (p = 0.038), via GA genotype. Even afteradjustment for sex, age, smoking status, alcohol consumption, presenceof diabetes, BMI and lipid-lowering drugs (Table 5), logistic regressionanalysis disclosed the APOA1−75GA genotype in men to exhibit signif-icantly elevated odds ratio for atherogenic dyslipidemia (OR = 1.57,95% Cl 1.06–2.3) compared with G homozygotes.

No significant relationship was found between the APOA1−75GNApolymorphism and MetS, CHD, diabetes, and obesity (BMI ≥ 30 kg/m2)in men and women (data not shown) using chi-squared test.

4. Discussion

This cross-sectional study is the first report to describe the genotypedistribution of the APOA1 −75GNA polymorphism in the TARF Study,representative of Turkish adults. Association of −75GNA polymor-phism in the 5′ untranslated region of the APOA1 gene with serumHDL-C concentrations, confirmed in the majority of studies [8–10,12],was also demonstrated among Turks, the −75GG genotype beingassociated with lower, the A allele inducing higher levels. Furtherrelationship between this SNP and several MetS components (such asatherogenic dyslipidemia, glucose concentrations and blood pressurelevels) was observed only in men.

The−75A allele frequency of the APOA1 −75GNA polymorphism inour population is similar to that previously reported in the Europeanpopulations (11.2–15.5%) [8–10] and Japanese population (13.5%) [25].However, the frequency of this allele is higher in some populations such

Table 3Adjusted mean (±SE) serum HDL-C, apoA-I, fasting glucose values, apoA1/HDL-C ratio and blood pressures stratified to gender and APOA1 genotypes.

Genotypes

N GG GA AA p*

Men (n = 696)HDL-cholesterol (mg/dL) 692 38.5 ± 0.50 38.3 ± 0.87 43.5 ± 2.08 0.057Fasting triglyceride (mg/dL)** 692 142.9 ± 1.02 145.2 ± 1.04 143.6 ± 1.10 0.943Apolipoprotein A-I (mg/dL) 400 129.5 ± 1.59 126.1 ± 2.82 136.8 ± 5.69 0.240apoA1/HDL-C ratio 400 3.56 ± 0.06 3.50 ± 0.10 3.01 ± 0.23 0.064Fasting glucose (mg/dL) 669 100.6 ± 1.03 95.7 ± 1.81 105.6 ± 4.21 0.021Systolic blood pressure (mm Hg) 692 127.1 ± 0.87 131.1 ± 1.51 123.2 ± 3.62 0.031Diastolic blood pressure (mm Hg) 692 80.0 ± 0.55 84.6 ± 0.96 80.8 ± 2.29 0.005

Women (n = 757)HDL-cholesterol (mg/dL) 756 45.5 ± 0.52 47.7 ± 0.86 52.2 ± 2.79 0.008Fasting triglyceride (mg/dL)** 756 121.9 ± 1.02 119.4 ± 1.03 107.4 ± 1.11 0.449Apolipoprotein A-I (mg/dL) 448 144.6 ± 1.65 148.4 ± 2.57 138.8 ± 9.81 0.378apoA1/HDL-C ratio 448 3.33 ± 0.05 3.23 ± 0.08 2.95 ± 0.29 0.285Fasting glucose (mg/dL) 737 98.2 ± 0.82 99.1 ± 1.34 91.7 ± 4.32 0.258Systolic blood pressure (mm Hg) 755 133.3 ± 0.99 134.9 ± 1.63 138.3 ± 5.25 0.476Diastolic blood pressure (mm Hg) 755 83.1 ± 0.57 83.6 ± 0.94 87.6 ± 3.03 0.329

Values were adjusted for age, waist circumference, current smoking and diabetes mellitus; lipid values additionally for lipid lowering medication, blood pressure values forantihypertensive drug usage and atherogenic dyslipidemia. **Geometric mean values. Significant values are highlighted in boldface. *p-values were obtained by ANCOVA for comparing−75GNA genotypes. The APOA1 −75GNA polymorphism had significant gender-by-genotype interaction (p = 0.017) in relation to glucose concentration, not to other parameters.

247N. Coban et al. / Clinica Chimica Acta 431 (2014) 244–249

as Taiwanese (32%) [26], Spanish and Omani (20.3–24%) [16,27,28]. It iswell recognized that ethnic differences exist in allele frequency betweendifferent populations.

Previous studies have reported that the APOA1/C3/A4/A5 gene clus-ter is associated with CHD and serum lipid levels [29,30]. Furthermore,polymorphisms in the APOA1 gene are associated with either increasedor decreased apoA1 and HDL-C concentrations [8–10,12,15]. Discre-pancies between these studies have been attributed to differences ingenetic susceptibility between different groups. It is possible that the as-sociation of APOA1 alleles and the apoA1 protein levels is influenced byother members of this cluster (APOC3, A4, and A5), which could be inlinkage disequilibrium in some populations but not in others [31]. Addi-tionally, such contradiction between studies may reflect differences inpopulation size, study design, gender, genetic heterogeneity or indeedthe dietary environment of the population study.

Unlike apoA1 levels, higher HDL-C concentration in our populationwas affected by APOA1 −75AA genotype as in most other reports[8–10,12]. The HDL-C is known to affect the atherogenic process andtherefore changes in HDL-C levels are correlated with prevalence ofatherosclerosis. Here, although APOA1−75GNA polymorphism is associ-ated with higher HDL-C levels, it seems not to be associated with CHD.Thismight be due to several reasons: One is, that the number of individ-uals with CHD in our study population is relatively low (111 cases in1515 people), the second reason might be the lack of measurement of

Table 4Frequencies of APOA1 −75GNA polymorphism in men with and without atherogenicdyslipidemia.

Variables With dyslipidemia Without dyslipidemia p-value

Genotype frequencya

GG, % (n) 68.1 (156) 74.1 (344)GA, % (n) 29.3 (67) 21 (100) 0.038AA, % (n) 2.6 (6) 4.9 (23)GA + AA, % (n) 31.9 (73) 26.3 (123) 0.127

Allele frequencyG, % 83 84.2 0.441A, % 17 15.8

a Frequency of alleles was computed using the gene-counting method. Chi-square testwas used for the association between the genotype frequencies and atherogenicdyslipidemia.

subclasses of HDL-C, such as preβ1,2 HDL and HDL2a, 2b, 3b. Many retro-spective and prospective studies have identified that small size HDLsubclasses contribute to increased risk of CHD while large size HDLsubclasses are associated with decreased risk [32]. It was postulatedthat HDL subclass distribution was altered and HDL particle size shiftedtowards smaller-sized HDL in subjects with GA mutation [33]. Thecorrelation between the genotypes and subclasses of HDL needs to beevaluated in a larger group with CHD.

This is the first study, to our knowledge, to report an interactionbetween gender and −75GA genotype of APOA1 with respect to an in-dependent risk for atherogenic dyslipidemia, inasmuch as only mendisplayed this susceptibility, after adjustment for conventional riskfactors. A significant association of APOA1 −75GNA polymorphismwas reported elsewhere with components of MetS such as HDL-C con-centrations [8–10,19], blood pressure levels [20], hypertension [18],BMI [17,19] and diabetes [16]. The genetic associations with MetS riskappeared in a French population sample to be modulated, interestingly,by gender, dietary fat intake and composition [19].

An association between APOA1−75GNA polymorphism and hyper-tension or blood pressure has been reported also elsewhere [18,20].Sadaf et al. found a significant association of APOA1 −75GA and AAgenotypes with higher systolic and diastolic BP in 162 Arab subjects[20]. In the current sample, the examined polymorphismwas similarly,significantly related to systolic and diastolic blood pressure, in malesalone. Male−75A allele carriers had significantly (by 3 mmHg) higherdiastolic BP thanmale G homozygotes. This is consistentwith a previousTARF study [34] in which serum apoA-I concentrations were linearlyrelated to systolic BP, independent of various confounders only inmen. Moreover, prehypertension was found to be mediated by apoA-Ilevels in each gender, in men (but not in women) being independentof serum triglycerides. We interpret the interaction between serumapoA-I and BP/prehypertension as a gene–environment interaction,the environmental factor being enhanced low-grade inflammation [34].

The cross-sectional design of the study is no more than a minorlimitation since genetic influences affect the individual lifelong. Ourfocusing on APOA1 −75GNA polymorphism, without assessing thecombined effect of other interacting genetic factors on analyses maybe a limitation to our study. The relatively large sample size and multi-variable adjustments to elicit independent effects form the strength ofthis study.

In conclusion, −75GNA polymorphism influences serum HDL-Clevels among Turkish adults by about 12–14%. Beyond this, gender-

Table 5Multiple logistic regression analysis showing the associations for atherogenic dyslipidemia in the total sample, men and women.

OR (95% CI) p OR (95% CI) P OR (95% CI) p

Total Men Women

Sex, female 0.53 (0.40–0.71) b0.001 – – – –

Age, years 0.99 (0.98–1.01) 0.25 0.98 (0.96–0.99) 0.006 1.01 (0.99–1.03) 0.22APOA1 −75 GA 1.26 (0.96–1.65) 0.101 1.57 (1.06–2.31) 0.023 0.94 (0.63–1.39) 0.76APOA1 −75 AA 0.5 (0.23 −1.07) 0.088 0.58 (0.22–1.51) 0.27 0.32 (0.07–1.45) 0.14Diabetes mellitus 1.54 (1.00–2.36) 0.049 1.09 (0.57–2.08) 0.79 2.43 (1.35–4.40) 0.003BMI (kg/m2) 1.11 (1.09–1.15) b0.001 1.16 (1.11–1.22) b0.001 1.09 (1.05–1.13) b0.001Lipid lowering drug usage 2.37 (1.31–4.30) 0.004 2.05 (0.76–5.52) 0.16 2.68 (1.26–5.73) 0.010Current smokers 1.32 (0.96–1.81) 0.089 1.34 (0.87–2.06) 0.19 1.32 (0.80–2.18) 0.28Alcohol consumption 0.85 (0.55–1.32) 0.47 0.76 (0.48–1.20) 0.24 2.82 (0.45–17.7) 0.27

Models comprised 1403 individuals (a missing value in 7% of participants) with 404 cases of atherogenic dyslipidemia (735 were women of whom 185 were cases of atherogenicdyslipidemia). Significant values are highlighted in boldface and italics.Referent for GA and AA genotypes was GG genotype. APOA1 −75GNA polymorphism had gender–genotype interaction (p= 0.001) in relation to atherogenic dyslipidemia.

248 N. Coban et al. / Clinica Chimica Acta 431 (2014) 244–249

specifically affects MetS traits, as atherogenic dyslipidemia and bloodpressure levels. Heterozygosity of this variant in men was associatedindependentlywith a 1.5-fold atherogenic dyslipidemia risk. The associ-ation of −75GA genotype with higher blood pressure levels in Turkishmen constitutes a genetic basis to prehypertension, documented previ-ously to be independently linked to apoA − I concentrations. Furtherstudies are warranted in different ethnicities to elucidate gene–gender–environment interactions.

Conflict of interest statement

The authors declare no competing interests.

Acknowledgment

This study was supported by the Research Support Unit of IstanbulUniversity as project no ACIP-4578. We thank the Turkish Society ofCardiology and several pharmaceutical companies, Istanbul, that havesupported financially the Turkish Adult Risk Factor survey. The partiallogistic support of the Turkish Ministry of Health is acknowledged.

References

[1] Expert panel on detection, evaluation and treatment of high blood cholesterol inadults (Adult Treatment Panel III). Executive summary of the Third Report of theNational Cholesterol Education Program (NCEP). JAMA 2001;285:2486–97.

[2] Gaziano JM, Hennekens CH, O'Donnell CJ, Breslow JL, Buring JE. Fasting triglycerides,high-density lipoprotein and the role of myocardial infarction. Circulation1997;96:2520–5.

[3] McLaughlin T, Abbasi F, Cheal K, Chu J, Lamendola C, Reaven G. Use of metabolicmarkers to identify overweight individuals who are insulin resistant. Ann InternMed 2003;139:802–9.

[4] Holleboom AG, Vergeer M, Hovingh GK, Kastelein JJ, Kuivenhoven JA. The value ofHDL genetics. Curr Opin Lipidol 2008;19:385–94.

[5] Teslovich TM, Musunuru K, Smith AV, et al. Biological, clinical and populationrelevance of 95 loci for blood lipids. Nature 2010;466:707–13.

[6] Rubin EM, Krauss RM, Spangler EA, Verstuyft JG, Clift SM. Inhibition of early athero-genesis in transgenic mice by human apolipoprotein AI. Nature 1991;353:265–7.

[7] Larson IA, Ordovas JM, Barnard JR, et al. Effects of apolipoprotein A-I genetic varia-tions on plasma apolipoprotein, serum lipoprotein and glucose levels. Clin Genet2002;61:176–84.

[8] Talmud PJ, Ye S, Humphries SE. Polymorphism in the promoter region of the apoli-poprotein AI gene associated with differences in apolipoprotein AI levels: theEuropean Atherosclerosis research study. Genet Epidemiol 1994;11:265–80.

[9] Jeenah M, Kessling A, Miller N, Humphries SE. G to A substitution in the promoterregion of the apolipoprotein AI gene is associated with elevated serum apolipopro-tein AI and high density lipoprotein cholesterol concentrations. Mol Biol Med1990;7:233–41.

[10] Sigurdsson GJ, Gudnason V, Sigurdsson G, Humphries SE. Interaction between apolymorphism of the apoA-I promoter region and smoking determines plasmalevels of HDL and apoA-I. Arterioscler Thromb 1992;12:1017–22.

[11] Wang XL, Badenhop RB, Sim AS, Wilcken DE. The effect on transcription efficiency ofthe apolipoprotein AI gene of DNA variants at the 5′ untranslated region. Int J ClinLab Res 1998;28:235–41.

[12] Pagani F, Sidoli A, Giudici GA, Barenghi L, Vergani C, Baralle FE. Human apolipopro-tein A-I gene promoter polymorphism: associationwith hyperalphalipoproteinemia.J Lipid Res 1990;31:1371–7.

[13] Barre DE, Guerra R, Verstraete R, Wang Z, Grundy SM, Cohen JC. Genetic analysis of apolymorphism in the human apolipoprotein A-I gene promoter: effect on plasmaHDL-cholesterol levels. J Lipid Res 1994;35:1292–6.

[14] de Franca E, Alves JG, Hutz MH. APOA1/C3/A4 gene cluster variability and lipid levelsin Brazilian children. Braz J Med Biol Res 2005;38:535–41.

[15] Xu CF, Angelico F, Del Ben M, Humphries SE. Role of genetic variation at the apoAI-CIII-AIV gene cluster in determining plasma apoA-I levels in boys and girls. GenetEpidemiol 1993;10:113–22.

[16] Morcillo S, Cardona F, Rojo-Martinez G, et al. Association between MspI polymor-phism of the APO AI gene and type 2 diabetes mellitus. Diabet Med 2005;22:782–8.

[17] Block R, Corsetti J, Goldenberg I, et al. The common apolipoprotein A-1 polymor-phism −75ANG is associated with ethnic differences in recurrent coronary eventsafter recovery from an acute myocardial infarction. Heart Int 2009;30 4:e8 [http://dx.doi.org/10.4081/hi.2009.e8].

[18] Chen ES, Mazzotti DR, Furuya TK, et al. Apolipoprotein A1 gene polymorphisms asrisk factors for hypertension and obesity. Clin Exp Med 2009;9:319–25.

[19] Phillips CM, Goumidi L, Bertrais S, et al. Gene-nutrient interactions and gender maymodulate the association between ApoA1 and ApoB gene polymorphisms and met-abolic syndrome risk. Atherosclerosis 2011;214:408–14.

[20] Sadaf A, Siddiqui S, Lestringant GG, Frossard PM. Apolipoprotein AI promoter variantin blood pressure determination. Clin Genet 2002;61:314–6.

[21] Onat A. Risk factors and cardiovascular disease in Turkey. Atherosclerosis2001;156:1–10.

[22] Genuth S, Alberti KG, Bennett P, et al. Expert committee on the diagnosis and classi-fication of diabetes mellitus. Follow-up report on the diagnosis of diabetes mellitus.Diabetes Care 2003;26:3160–7.

[23] Grundy SM, Brewer HB, Cleeman JI, Smith SC, Lenfant C. Definition of metabolicsyndrome: report of the National Heart, Lung, and Blood Institute/American HeartAssociation conference on scientific issues related to definition. Circulation2004;109:433–8.

[24] Onat A, Uyarel H, Hergenc G, Karabulut A, Albayrak S, Can G. Determinants anddefinition of abdominal obesity as related to risk of diabetes, metabolic syndromeand coronary disease in Turkish men: a prospective cohort study. Atherosclerosis2007;191:182–90.

[25] Bai H, Saku K, Liu R, Jimi S, Arakawa K. Analysis of a new polymorphism in thehuman apolipoprotein A-I gene: association with serum lipoprotein levels andcoronary heart disease. J Cardiol 1996;28:207–12.

[26] Wu JH, Kao JT, Wen MS, Lo SK. DNA polymorphisms at the apolipoprotein A1-CIIIloci in Taiwanese: correlation of plasma APOCIII with triglyceride level and bodymass index. J Formos Med Assoc 2000;99:367–74.

[27] Lahoz C, Pena R, Mostaza JM, et al. Apo A-I promoter polymorphism influencesbasal HDL-cholesterol and its response to pravastatin therapy. Atherosclerosis2003;168:289–95.

[28] Al-Yahyaee SA, Al-Kindi MN, Al-Bahrani AI. Apolipoprotein A1 gene polymorphismsat the−75 bp and +83/84 bp polymorphic sites in healthy Omanis compared withworld populations. Hum Biol 2004;76:307–12.

[29] Yin RX, Li YY, Lai CQ. Apolipoprotein A1/C3/A5 haplotypes and serum lipid levels.Lipids Health Dis 2011;19:10–140 [http://www.lipidworld.com/content/10/1/140].

[30] Ding Y, Zhu MA, Wang ZX, Zhu J, Feng JB, Li DS. Associations of polymorphisms inthe apolipoprotein APOA1-C3-A5 gene cluster with acute coronary syndrome. JBiomed Biotechnol 2012. http://dx.doi.org/10.1155/2012/509420.

[31] Paul-Hayase H, Rosseneu M, Robinson D, Van-Bervliet JP, Deslypere JP, HumphriesSE. Polymorphisms in the apolipoprotein (apo) AICIII-AIV gene cluster: detectionof genetic variation determining plasma apo AI, apo CIII and apo AIV concentrations.Hum Genet 1992;88:439–46.

[32] Superko HR. Lipoprotein subclasses and atherosclerosis. Front Biosci 2001;6:355–65.[33] Jia L, Bai H, Fu M, Xu Y, Yang Y, Long S. Relationship between plasma HDL subclasses

distribution and apoA-I gene polymorphisms. Clin Chim Acta 2005;360:37–45.[34] Onat A, Örnek E, Çiçek G, Murat SN, Can G. Increased apolipoprotein A-I levels me-

diate the development of prehypertension. Anadolu Kardiyol Derg 2013;13:306–14.

N. Coban et al. / Clinica Chimica

NeslihanCoban is a biologist, graduated in 2003 inbiologicalsciences from Istanbul University, Turkey, a M.Sc. and a Ph.D.in genetics in 2012 from the Health Science Institute atIstanbul University, Turkey. She has been continuing herPh.D. education at the same institute and working as aresearch assistant at the Department of Genetics, Institute

for Experimental Medicine, Istanbul University. Her mainareas of interest in research are genetics and molecularepidemiology for cardiovascular diseases.

Altan Onat, cardiologist, scientist, medical educator; b. 1929.M.D., U. Zurich, 1954; Fellow in Medicine Mayo Clinic,Rochester, Minn., retired as prof. medicine from CerrahpaşaMed. Faculty, U. İstanbul, Turkey, 1996,. Conducted the ongo-ing Turkish Adult Risk Factor Survey (1990-). Mem. (2007-) ofthe study group investigating the epidemiology of diabetesin Europe andAsia. Takes part in themetaanalyses on Emerg-ing Risk Factors Collaboration (Cambridge U., UK). Receivedover 1500 citations per Science Citation Index. Served over100 times as reviewer of submitted manuscripts for over ascore of leadingmedical periodicals. Contributed 100 articlesto professional journals.

Filiz Geyik is a biologist, graduated from Istanbul University(Turkey) Biological SciencesDepartment in 2001, aM.Sc. anda Ph.D. in genetics in 2012 from the Health Science Instituteat Istanbul University, Turkey. Since 2005, she has beenworking as a research assistant at the Department of Genet-ics, Institute for Experimental Medicine, Istanbul University.Her research interests are genetics, molecular epidemiologyand epigenetics for cardiovascular diseases.

Evrim Komurcu-Bayrak received a B.S. in medical biologyfrom the Cerrahpasa Medical Faculty in 1998, a M.Sc. and aPh.D. in genetics in 2008 from the Health Science Instituteat Istanbul University, Turkey. As a research assistant, shebegan to work at the Department of Genetics, Institute forExperimental Medicine, Istanbul University in 1998, andjoined several collaborative studies in site visits. Her mainareas of interest in research are genetics, molecular epidemi-ology and candidate gene mutation screening for heartdiseases.

249Acta 431 (2014) 244–249

Günay Can graduated at Cerrahpasa Medical School ofIstanbul University. He has been teaching in the Departmentof Public Health at the Cerrahpasa Medical Faculty, IstanbulUniversity. He is a statistician and works for statisticalanalysis of TARF (Turkish Adult Risk Factor) study.

Nihan Erginel-Unaltuna completed basic medical sciencesfrom the School of Medicine, University of Vienna, Austria1987 and received a M.Sc. (1991) and a Ph.D. (1994) ingenetics from Cell and Molecular Biology the SUNY HealthScience Center, Syracuse, NY. She became full professor of ge-netics in 2002 at theDepartment of Genetics, Institute for Ex-perimental Medicine, Istanbul University and has beenworking in the same institute. Her main areas of interest inresearch are genetics, molecular cardiology, metabolic syn-drome and neurogenetics.