Association of gender, ABCA1 gene polymorphisms and lipid profile in Greek young nurses

Objective One of the important proteins involved in lipid metabolism is the ATP-binding cassette transporter A1 (ABCA1) encoding by ABCA1 gene. In this study we evaluated the single nucleotide polymorphisms (SNPs) of ABCA1 gene. We analyzed SNPs in chromosome 9 such as rs2230806 (R219K) in the position 107620867, rs2230808 (R1587K) in the position 106602625 and rs4149313 (I883M) in the position 106626574 according to gender and lipid profile of Greek nurses. Methods The study population consisted of 447 (87 men) unrelated nurses who were genotyped for ABCA1 gene polymorphisms. Additionally, lipid profile [total cholesterol, triglycerides, high density lipoprotein cholesterol, low density lipoprotein cholesterol (LDL-C) and apolipoprotein A1] was evaluated. Results The distribution of all three studied ABCA1 gene polymorphisms did not differ according to gender. However, only R219K genotype distribution bared borderline statistical significance (p = 0.08) between the two studied groups. Moreover, allele frequencies of R219K, R1587K and I88M polymorphisms did not differ according to gender. In general, blood lipid levels did not seem to vary according to ABCA1 gene polymorphisms, when testing all subjects or when testing only men or only women. However, a significant difference of LDL-C distribution was detected in all subjects according to R1587K genotype, indicating lower LDL-C levels with KK polymorphism (p = 0.0025). The above difference was solely detected on female population (p = 0.0053). Conclusions The ABCA1 gene polymorphisms frequency, distribution and lipid profile did not differ according to gender. However, in the female population the KK genotype of R1587K gene indicated lower LDL-C levels. Further studies, involving a higher number of individuals, are required to clarify genes and gender contribution.


Introduction
ATP-binding cassette transporter A1 (ABCA1) mediates the transport of cholesterol and phospholipids from cells to lipid-poor apolipoproteins. Animals and human studies documented that defects in the ABCA1 pathway are significant determinants of coronary artery disease (CAD) [1]. Inactivation of ABCA1 gene in macrophages increases atherosclerotic lesions in hyperlipidemic mice [2,3], and overexpressing human ABCA1 in transgenic mice retards atherogenesis [4,5]. The ABCA1 gene is located on the chromosome 9 in the area 9q31.1 and encodes ABCA1 protein. ABCA1 protein is expressed in liver, macrophages, intestines, lungs etc. Several ABCA1 gene polymorphisms were identified such as rs2230806 (R219K) in the chromosomal position 107620867, rs2230808 (R1587K) in the chromosomal position 106602625 and rs4149313 (I883M) in the chromosomal position 106626574]. This study in line with our previous work [6] was undergone to evaluate the association of gender, three ABCA1 gene polymorphisms mention above and lipid profile in Greek male and female nurses.

Subjects
The genotyping of 447 (87 male) nurse students median age 22 (21)(22)(23)(24)(25) years old, who were attended to the University of Nursing of Technological and Educational Institution, was performed. All students had no personal history of CAD and were not taking any drugs. Other exclusion criteria were diabetes mellitus, thyroid and liver disease, high alcohol consumption, professional athleticism and any chronic disease.
All students were attended to the University every day and were staying for 8-10 hours. Students were eating (breakfast, snacks and lunch) at the school canteen which served typical Mediterranean food. Only one meal daily (dinner) was most likely to be different in each student.
The University of Nursing of Technological and Educational Institution Ethics Committee approved the protocol of this study. All subjects signed an informed consent form.

Blood chemistry
Plasma total cholesterol (TC), triglycerides (TGs), high density lipoprotein cholesterol (HDL-C) and apolipoprotein A1 (Apo A1) were measured using enzymatic colorimetric methods on Roche Integra Biochemical analyzer with commercially available kits (Roche). The serum low density lipoprotein cholesterol (LDL-C) concentration was calculated using the Friedewald formula only in subjects with TGs concentration < 400 mg/dl.

DNA analysis and determination of blood lipids
The ABCA1 gene polymorphisms (R219K, R1587K and I883M) were detected using polymerase chain reaction (PCR) and restricted fragment length polymorphism analysis (RFLP's). The PCR was performed using Taq polymerase KAPATaq. The oligonucleotide primers used for R219K and R1587K polymorphisms were described by Saleheen D et al. [7] and Tupitsina TV et al. [8] respectively.
The oligonucleotide primers used for I883M polymorphism were 5'-GAGAAGAGCCACCCTGGTT-CCAACCAGAAGAGGAT-3' and 5'-AGAAAGGCAG-GAGACATCGCTT −3 as described by Clee SM et al. [9]. PCR was subjected to 95 ο C for 5 min, thirty cycles of 95 ο C for 30 s, 65 ο C for 30 s and 72 ο C for 30 s and final extension to 72 ο C for 7 min, producing a fragment of 132 bp. This fragment was subsequently cleaved by EcorV, creating fragments for I allele 97 bp and 35 bp and for M allele 132 bp, which were subjected to electrophoresis on an agarose gel 4% and visualized with ethidium bromide (Figure 1).

Statistical analysis
The Shapiro-Wilk test was performed to test for normal distribution of continuous variables. The results are given as median and interquartile range (IQR), whereas, all qualitative variables are presented as absolute or relative frequencies. The Mann-Whitney U test or the Fisher's exact test was employed for comparison of continuous or categorical variables, respectively. The comparison within groups was performed by test for equality of proportions (Bonferroni correction). The Kruskal -Wallis H statistic was employed in order to detect differences in lipid levels according to three different polymorphisms of ABCA1 gene (R219K, R1587K and I883M). ABCA1 polymorphisms' distribution according to lipid profile was tested using the Kruskal Wallis test in men, in women and in all patients. All results were corrected for multiple testing (Bonferroni correction). All tests were two-tailed and statistical significance was established at 5% (p < 0.05). Data were analysed using Stata ™ (Version 10.1 MP, Stata Corporation, College Station, TX 77845, USA).

Results
Comparison of various characteristics according to gender showed that the study group was homogenous regarding age (p > 0.05) ( Table 1). However, significant differences were detected regarding other anthropometric characteristics and lipidemic profile (Table 1). Specifically, BMI, waist and mean blood pressure measurements were higher in men when compared to women (Table 1). Moreover, TG levels were significantly elevated in men, whereas HDL-C and Apo A1 levels were significantly elevated in women. No differences were detected for TC measurements between men and women ( Table 1).
The distribution of all ABCA1 gene polymorphisms studied did not differ according to gender (Table 2). However, only R219K distribution bared borderline significance between the two groups studied ( Table 2). A post hoc power calculation was conducted employing the calculated differences in the distribution of R219K polymorphism between men and women. The post hoc power (<0.6) was not sufficient enough to support the hypothesis. Furthermore, none of the comparisons within groups (concerning the rest of the polymorphisms) were statistically significant (p > 0.1). Moreover, allele frequencies of R219K, R1587K and I88M polymorphisms did not differ according to gender (Table 2).
In general, blood lipid levels did not seem to vary according to ABCA1 gene polymorphisms, when testing all subjects (Table 3) or when testing only men (Table 4) or only women (Table 5). However, a significant difference of LDL-C distribution was detected in all patients according to R1587K genotype, indicating lower LDL-C levels with KK polymorphism ( Table 3). The above difference was solely detected on female population (Table 5).

Discussion
In this study we evaluated the association of gender, ABCA1 gene polymorphisms and lipid profile in Greek young nurses.
In previous study we evaluated the influence of ABCA1 gene polymorphisms on lipid profile in female Greek nurses. In the present study, we assessed more subjects, included men and evaluated one more polymorphism, namely rs4149313 (I883M).

R219K polymorphism
One of the first study which evaluated the R219K polymorphism and lipid profile in man was published by Clee SM et al. [9]. They genotyped 804 Dutch men with CAD who were participated in the Regression Growth Evaluation Statin Study (REGRESS). The frequency of K allele in their study was similar to our study (25.4% and 28%, respectively). They found that K allele carriers had decreased progression of atherosclerosis and a reduced risk of coronary events. Furthermore, the K allele was associated with lower plasma TGs concentration and a trend toward higher HDL-C concentration. Since then, many other studies correlated the R219K variant with the lipid profile and risk for CAD in various populations and ethnicities [6,[9][10][11][12][13][14][15][16]. Conversely, Pasdar et al. [10] did not support a major role of the ABCA1 gene as risk factor for ischemic stroke. Srinivasan et al. [11] found that the K219 allele frequency differs markedly between blacks and whites, and that the variant-allele modulates the association between age and HDL-C, as well as body fatness and TGs concentration in a beneficial manner only in whites. Bertolini et al. [12] found that in subjects with CAD, the prevalence of RK and KK genotypes was lower than in subjects free of CAD (33.0% versus 51.5%). Zhao et al. [13] found a significant upward  trend in the sequence of RR, RK, and KK genotypes with a significant difference between RR genotype (50.5 ± 15.6 mg/dl, 1.3 mmol/l ± 0.4 mmol/l) and KK genotype (54.6 ± 15.6 mg/dl, 1.4 mmol/l ± 0.4 mmol/l), especially in males. This was not proved by our previous study where only women were participated [6]. Additionally, Frikke-Schmidt et al. [14] did not find any association with HDL-C levels in Danish population.
Kitjaroentham et al. [15] studied overweight/obese Thai subjects of both sexes. The overweight/obese men who had the K allele had lower HDL-C than controls. In our study, we do not find any association of the R219K variant with any parameter of the lipid profile. Also, the distribution of all studied ABCA1 gene polymorphisms was not different according to gender. Noteworthy to mention is that the R219K distribution bared borderline statistical significance (p = 0.08) between two studied groups. In general, blood lipid levels did not seem to vary according to ABCA1 gene polymorphisms, when testing all subjects or when testing only men or only women.

R1587K polymorphism
Clee et al. [9] in the study already mention above found that K carriers of R1587K (RK, KK) had lower HDL-C concentration compared with noncarriers in an allele dose-dependent trend. On multiple regression analysis including age, BMI, smoking, and TG as covariates, the R1587K genotype remained a significant predictor of HDL-C. Frikke-Schmidt et al. [16] associated the R1587K variant with a stepwise decrease in HDL-C in women in heterozygotes and homozygotes, and with a similar trend in men. Mantaring et al. [17] did not find any statistical significance between R1587K variant with the lipid profile. However, Tsai et al. [18] found that after fenofibrate treatment the KK genotype of R1587K was associated with significantly increased small HDL and with increased HDL particle concentrations. In our previous study [6] where only women were participated, the difference in TC, LDL-C and TGs concentration was detected between RK and RR genotypes. In this study we found according to gender that blood lipid levels did not  seem to vary in genotypes of R1587K polymorphism. However, a significant difference of LDL-C distribution was detected in female population. We also expected to found a difference in TC and TGs values between genotypes of female population as in previous study [6].
However, in the current study, in order to expand the study population, we included women with different lipid profiles, affecting the overall variability and altering the final results. Generally, lipid profile can be influenced by several genetic and environmental factors such as smoking status, eating habits, associated customs (fasting periods), sedentary life style, body composition, gender and others. The most important factors influencing the lipid profile include the sex hormone status and age. Unfortunately, in our study apart from age, smoking status and BMI we have not gone to all other details which potentially could influenced the results (added data consisted of 1/6 of women's study cohort).

I883M polymorphism
Tan et al. [19] studied the I883M variant in Malays and Chinese population. They found an association with lipoprotein(a) concentration in Malays population and with apolipoprotein B concentration in Chinese population. However, Clee et al. [9] did not find any difference according to lipid levels and genotypes in carriers of the I883M, although individuals with MM genotype had higher progression in minimum obstruction diameter and cardiac event rate compared with the II genotype individuals. Hodoğlugil et al. [20] correlated the I883M variant with higher HDL-C concentration in both sexes. Similarly, Jensen et al. [21] among younger women and Porchay-Baldérelli et al. [22] in population with type 2 diabetes mellitus found that the M allele of I883M was associated with higher HDL-C concentration. Additionally, Mantaring et al. [17] found some differences in allele frequency of I883M gene, which were also present between the highest and lowest HDL-C concentration groups (36% vs 20%; P trend = 0.05). On the contrary, Kitjaroentham et al. [15] did not find any difference in HDL-C concentrations among I883M genotype polymorphism. Although, this variant was common in Thai  ethnic groups, there were no differences detected in genotype frequency between overweight/obese and control subjects groups. Similarly, Frikke-Schmidt et al. [14] found that I883M did not affect HDL-C levels but predicted higher risk of CAD. Also, Sandhofer et al. [23] found that I883M variant showed no effect on plasma lipids or carotid atherosclerosis. In our study, we have not found any association with the lipid profile. However, there are some data from Delgado-Lista et al. [24] suggesting that the major allele homozygotes for ABCA1 gene polymorphisms i27943 (rs2575875) and R219K (rs2230806) have a lower postprandial response as compared to minor allele carriers. The fact that in current study blood lipid levels did not seem to vary according to ABCA1 gene polymorphism could be influence, at least in part, because they were measured in the fasting state. A limitation of this study is the relatively small number of men's group. However, the effort was put for sample to be homogenous, living in the similar conditions as it happened with our study population.
Another limitation of this type of study is that, studies based on the candidate-gene approach, which have been demonstrated genotype-phenotype associations, are not always replicable.
There are only few studies which compare the influence of ABCA1 gene polymorphisms on lipid profile in accordance to gender. The ABCA1 gene polymorphisms frequency, distribution and lipid profile did not differ according to gender. However, in the female population the KK genotype of R1587K gene indicated lower LDL-C levels.
genetic studies, AK performed the statistical analysis and drafting of the manuscript, GV and AK collected the blood samples, SM participated in the drafting and revising the manuscript, DD participated in revising the manuscript critically for important intellectual content, CM participated in the study design and its coordination and GK conceived the study and participated in the development of the hypothesis, the study design and drafting of the manuscript. All authors read and approved the final manuscript.