The SRGAP2 SNPs, their haplotypes and G × E interactions on serum lipid traits

Maonan nationality is a relatively conservative and isolated minority in China. Little is known about the association of the Slit-Robo Rho GTPase activating protein 2 gene (SRGAP2) single nucleotide polymorphisms (SNPs) and serum lipid levels in the Chinese populations. This study was performed to clarify the association of the SRGAP2 rs2483058 and rs2580520 SNPs and their haplotypes with serum lipid traits in the Maonan and Han populations. Genotyping of the 2 SNPs was performed in 2444 unrelated subjects (Han, 1210 and Maonan, 1234) by polymerase chain reaction and restriction fragment length polymorphism combined with gel electrophoresis, and then confirmed by direct sequencing. The allelic (rs2483058) and genotypic (rs2483058 and rs2580520) frequencies were different between the two ethnic groups. Four haplotypes were identified in our populations, and the rs2483058G-rs2580520C haplotype was the commonest one. The rs2483058C-rs2580520G haplotype was associated with an increased risk of dyslipidemia, and showed consistent association with serum total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), apolipoprotein (Apo) A1 levels, and the ApoA1/ApoB ratio. These results indicated that the SRGAP2 SNPs and their haplotypes were associated with serum lipid levels. Their haplotypes can explain much more serum lipid variation than any single SNP alone, especially for serum TC, HDL-C and ApoA1 levels.


Results
Demographic and biochemical characteristics. The demographic and biochemical characteristics of the participants according to ethnic group are presented in Table 1. The levels of body weight, body mass index (BMI), waist circumference, systolic blood pressure, diastolic blood pressure, pulse pressure and blood glucose and the percentages of subjects who consumed alcohol were higher in Maonan than in Han (P < 0.05-P < 0.001), whereas the levels of body height and the percentages of subjects who smoked cigarettes were lower in Maonan than in Han (P < 0.05). There was no significant difference in the levels of age, sex ratio, TC, TG, LDL-C, HDL-C, ApoA1, ApoB and the ratio of ApoA1 to ApoB (P > 0.05 for all).  Table 1. Comparison of demographic, lifestyle characteristics and serum lipid levels between the Han and Maonan populations. HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. 1 The quantitative variables were presented as mean ± standard deviation and determined by Student's unpaired t-test. 2 The value of triglyceride was presented as median (interquartile range) for not meet the normal distribution, the difference between the two ethnic groups was determined by the Wilcoxon-Mann-Whitney test.

Results of electrophoresis and genotyping.
After the genomic DNA of the samples was amplified by polymerase chain reaction (PCR) and imaged by agarose gel electrophoresis for the SRGAP2 rs2483058 SNP, the PCR product of 382-bp nucleotide sequences could be seen in the samples (Figs 1 and 2). The GG (382 bp), GC (382-, 260-and 122-bp) and CC (260-and 122-bp) genotypes were shown, respectively. The PCR product of the rs2580520 SNP was 515-bp nucleotide sequences. The GG (515 bp), GC (515-, 389-, and 126-bp) and CC (389and 126-bp) genotypes were shown, respectively.
Genotypic and allelic frequencies. The genotypic distribution of the two loci was in Hardy-Weinberg equilibrium (P > 0.05 for all). The genotypic and allelic frequencies of the rs2483058 SNP were different between Maonan and Han, the frequencies of the rs2483058C allele (41.86% vs. 37.07%, P = 0.001) and rs2483058GC genotype (48.06% vs. 40.90%, P < 0.001) were higher in Maonan than in Han. Subgroup analysis showed that the rs2483058C allele frequency was higher in Han females than in Han males (39.06% vs. 33.87%, P = 0.01), but it was lower in Maonan females than in Maonan males (39.37% vs. 45.50%, P = 0.002). The genotypic frequencies of the rs2580520 SNP between Maonan and Han, and between Han males and Han females were also different (P < 0.05-P < 0.005; Table 2). Tables 3 and 4 describe the association between genotypes and serum lipid levels. The levels of ApoA1 and the ratio of ApoA1 to ApoB in both Han and Maonan were different among the rs2483058 genotypes (P < 0.01 for all), the rs2483058C allele carriers had lower ApoA1 levels and the ApoA1/ ApoB ratio than the rs2483058C allele non-carriers. The concentration of HDL-C in Maonan but not in Han was also different among the rs2483058 genotypes (P = 0.024), the rs2483058C allele carriers had lower HDL-C concentration than the rs2483058C allele non-carriers. The levels of TC, LDL-C and ApoB in Maonan but not in Han were different among the rs2580520 genotypes (P = 0.011-P = 0.001), the rs2580520C allele carriers had higher TC, LDL-C and ApoB levels than the rs2580520C allele non-carriers. Subgroup analysis showed that the levels of ApoA1 and the ratio of ApoA1 to ApoB in Han males and females and Maonan males were different among the rs2483058 genotypes (P < 0.01 for all), the rs2483058C allele carriers had lower ApoA1 levels and the ApoA1/ApoB ratio than the rs2483058C allele non-carriers. The levels of ApoA1 in Maonan females were also different among the rs2483058 genotypes (P < 0.001), the rs2483058C allele carriers had lower ApoA1 levels than the rs2483058C allele non-carriers. The levels of TC in Maonan males and females were different among the rs2580520 genotypes (P = 0.018-P = 0.001), the rs2580520C allele carriers had higher TC levels than the rs2580520C allele non-carriers. The ratio of ApoA1 to ApoB in Han males and Maonan males was also different among the rs2580520 genotypes (P = 0.013-P = 0.007), the rs2580520C allele carriers had lower the ApoA1/ApoB ratio than the rs2580520C allele non-carriers.  Haplotypes and serum lipid levels. A weak linkage disequilibrium (LD) was noted between the two SNPs (D′ = 0.58, r 2 = 0.33; Fig. 3). Thus, a haplotype analysis was conducted (Table 5). There were 4 haplotypes identified in our study populations. The haplotype of rs2483058G-rs2580520C was the commonest one (45%). The haplotype of rs2483058C-rs2580520G was associated with an increased risk of dyslipidemia (OR: 1.44, 95% CI: 1.17-1.78, P < 0.001), it showed consistent association with serum TC, HDL-C, ApoA1 and the ratio of ApoA1 to ApoB. Multivariate logistic analysis showed that the rs2483058C-rs2580520G haplotype was positively correlated with the incidence of hyperlipidemia in Han and Maonan according to stratified risk factors (gender, BMI, smoking, diabetes and blood pressure; Table 6). In addition, the rs2483058C-rs2580520G haplotype carriers had lower HDL-C and ApoA1 levels in Han and Maonan, lower ApoA1/ApoB ratio in Han, and higher TC in Maonan than the rs2483058C-rs2580520G haplotype non-carriers (Table 7).

Correlated factors for serum lipid parameters.
Multivariable linear regression analyses showed that the levels of HDL-C and ApoA1 were associated with the rs2483058 genotypes and/or alleles in Han. The levels of TC were correlated with the rs2580520 genotypes and alleles, HDL-C and ApoA1 with the rs2483058 genotypes in Maonan (Table 8).
Serum lipid parameters were also correlated with several environmental factors such as sex, age, alcohol consumption, cigarette smoking, blood pressure, blood glucose, waist circumference, and BMI in both ethnic groups or in males and females (P < 0.05-P < 0.001; Tables 9 and 10). Fig. 4, Pearson correlation analysis demonstrated that the integrative variants and haplotype connected with the SRGAP2 rs2483058 and rs2580520 SNPs to lipid variables. Several environmental exposures such as age, gender, cigarette smoking, alcohol consumption and traditional cardiovascular risk factors such as BMI and blood pressure levels were also correlated with serum lipid phenotypes in the both ethnic groups.

Discussion
The present study demonstrated that different SRGAP2 SNPs had different effect on serum lipid traits that (i) the levels of HDL-C, ApoA1 and the ratio of ApoA1 to ApoB in Maonan were different among the rs2483058 genotypes; (ii) the levels of TC in Maonan males and females and the ratio of ApoA1 to ApoB in Han and the level of LDL-C and ApoB in Maonan were different among the three rs2580520 genotypes; (iii) the levels of HDL-C and ApoA1 in males or females were different in Maonan and Han in rs2483058 genotypes; and (iv) the levels of TC and the ratio of ApoA1 to ApoB in males or females were different in Maonan and Han in rs2580520 genotypes.
In the current study, we found that the rs2483058C allele carriers had lower ApoA1 levels and the ApoA1/ ApoB ratio than the rs2483058C allele non-carriers in both ethnic groups. The rs2483058C allele carriers also had lower HDL-C concentration than the rs2483058C allele non-carriers in Maonan but not in Han. The rs2580520C allele carriers had higher TC, LDL-C and ApoB levels than the rs2580520C allele non-carriers in Maonan but not in Han. The levels of TG, LDL-C and ApoB in the two ethnic groups were not significantly different among the rs2483058 and rs2580520 genotypes. It is well-known that dyslipidemia is a complex disease caused by environmental and genetic factors. Previous family and twin studies have suggested that in numerous populations, about 40-60% of the variation in serum lipid profiles is genetically determined [32][33][34][35][36] .
In some previous studies, there were associations with intra-pair differences in HDL-C, rs2483058 in an intron of SRGAP2, where twins carrying the C allele were more sensitive to environmental factors 37 . In the present study, we showed that the genotypic frequencies of our populations were similar to those obtained in other populations and to data available in the International HapMap Project's database (http://www.hapmap.org) for the populations. The frequency of CC, GC and GG genotypes was 42.9%, 42.9% and 14.2% in Utah residents with ancestry from Northern and Western Europe (CEU); 7.2%, 56.4% and 36.4% in Yoruba in Ibadan, Nigeria (YRI); 9.3%, 37.2% and 53.4% in Japanese in Tokyo, Japan (JPT) and 9.5%, 40.5% and 50% in Han Chinese in Beijing, China (HCB). The frequency of C and G alleles was 64.3% and 35.7% in CEU; 27.9% and 72.1% in YRI; 35.5% and 64.5% in JPT and 29.8% and 70.2% in HCB. The present study identified significant differences in the genotypic frequency of the rs2483058 SNP between the two ethnic groups. The frequency of the CC genotype was higher in the Maonan than in the Han population and the frequency of the C allele was higher in the Maonan than in the Han population (41.86% vs. 37.03%; P = 0.001). These results suggest that the prevalence of the SRGAP2 rs2483058 SNP may exhibit a racial/ethnic difference.
In some previous studies, Jiang et al. had found that the rs2580520 genotypes in Chinese women may have different breast cancer susceptibility, which may contribute to the development of breast cancer in this population 27  five decades and the contribution of obesity to postmenopausal breast cancer risk is thought to be related to elevated level of circulating estrogens in these women, resulting from aromatase-mediated conversion of androgens to estrogens in peripheral adipose tissue (and possible dysregulation of aromatase expression) and lower level of sex hormone-binding globin 38 . It is well-known that cholesterol is an important material for the synthesis of estrogen and higher cholesterol can give rise to more estrogen synthetic that may be enhanced a possibility to be suffered from breast cancer 39,40 . The results of the present study clearly showed that the levels of TC were higher in who carried rs2580520C allele than without carried rs2580520C allele. Important intra-and inter-genetic LD associations have been found in this study. These LD patterns in SRGAP2 were rather weak and low specific to the population under study and indicated the functional dependencies of the encoded proteins. In the present study, haplotype analysis with the two SNPs further supported the association between SRGAP2 polymorphisms and serum lipid levels in our study populations. The haplotype of rs2483058G-rs2580520C was the commonest one and represented about 45% of the samples. The haplotypes of rs2483058C-rs2580520C, rs2483058G-rs2580520G and rs2483058C-rs2580520G were associated with increased risk of dyslipidemia, but only rs2483058C-rs2580520G haplotype with a significant meaning. In addition, carriers of rs2483058C-rs2580520G haplotype had decreased serum concentration of HDL-C, ApoA1 and the ratio of ApoA1 to ApoB and increased serum concentration of TC. We also found that haplotypes could explain much more serum lipid variation than any single SNP alone.
The environmental factors just as dietary patterns, lifestyle, obesity, physical activity, and hypertension would also be play an important role in modifying serum lipid levels [41][42][43] . The dietary habits were different between the Han and Maonan populations. Rice is the Maonan people's staple food supplemented with corn, sweet potato and other grains. Maonan people preferred to eat spicy and acid food with lots of oil and salt. This preference of high in carbohydrates may be related to the higher blood glucose levels, weight, BMI and waist circumference in Maonan than in Han people. In the meantime, rich oil and salt can give rise to higher blood pressure, serum TC, LDL-C and ApoB levels in Maonan than in Han people. There were lots of past studies had proved that diet alone could account for the variability on serum lipid levels 44,45 . In the present study, we also found that serum lipid     parameters were also correlated with several environmental factors such as age, alcohol consumption, cigarette smoking, blood pressure, blood glucose, waist circumference, and BMI. It is commonly accepted that the high-fat diet especially containing large quantities of saturated fatty acids, raise serum cholesterol concentrations and predispose subjects to coronary artery disease 46 . We also showed that the percentages of individuals who consumed alcohol were higher in Maonan than in Han. Although the effects of alcohol intake on LDL-C appear to vary by specific patient types or patterns of alcohol intake, and perhaps by population and sex hormone, this topic has been the focus of much recent research 47 . A recent study in older Italian individuals (65-84 years old) has found    that alcohol intake reached higher serum LDL-C levels 48 . Another recent study of Turks also found increased LDL-C, ApoB and TG with alcohol in men, whereas decreased TG and did not change LDL-C or ApoB with alcohol in women 49 . Therefore, the results of exposure to different lifestyle and environmental factors probably further modify the association of genetic variations and serum lipid levels in our study populations.
The present study has some shortcomings. At first, the size of our study populations is a bit small, which might not have the power to detect the LD across the SRGAP2 locus. Next, the levels of body weight, waist circumference and the percentages of subjects who smoked cigarettes or consumed alcohol were higher in Maonan than in Han. Although age, BMI, blood pressure, cigarette smoking, and alcohol consumption have been adjusted for the statistical analysis, we cannot completely exclude the influence of these factors on serum lipid levels among different genotypes in both nationalities. In addition, because we selected the SNPs from literature and did not cover the extensive SRGAP2 locus, we might miss some information from other SNPs.    In summary, the genotypic and allelic frequencies of the SRGAP2 rs2483058 and rs2580520 SNPs were different between Maonan and Han. There were four haplotypes identified in our study populations. The SRGAP2 SNPs and rs2483058C-rs2580520G haplotype were closely sexually dimorphic associated with serum lipid traits. The haplotypes can explain much more serum lipid variation than any single SNP alone, especially for serum TC, HDL-C and ApoA1 levels.

Materials and Methods
Subjects. The study populations including 1210 unrelated subjects (465 males, 38.43% and 745 females, 61.57%) of Han and 1234 unrelated participants (500 males, 40.52% and 734 females, 59.48%) of Maonan were randomly selected from our previous stratified randomized samples. The participants were all agricultural workers from Huanjiang County, Guangxi Zhuang Autonomous region, People's Republic of China. The participants' age ranged from 25 to 80 years with a mean age of 55.77 ± 13.89 years in Han and 56.31 ± 13.80 years in Maonan; respectively. The age distribution and gender ratio were matched between the two groups. All participants were essentially healthy with no history of CVD such as coronary artery disease, stroke, diabetes, hyper-or hypo-thyroids, and chronic renal disease. They were free from medications known to affect serum lipid levels. The investigations were carried out following the rules of the Declaration of Helsinki of 1975 (http://www.wma.net/ en/30publications/10policies/b3/), revised in 2008. The Ethics Committee of the First Affiliated Hospital, Guangxi Medical University approved the study protocol (No: Lunshen-2014-KY-Guoji-001; Mar. 7, 2014) prior to data collection, and all participants provided informed consent by signature or by fingerprint (if the participant was illiterate) after they had been informed of the objectives, benefits, medical items and confidentiality agreement of personal information. All procedures and methods were performed in accordance with the relevant ethical guidelines.
Epidemiological survey. The epidemiological survey was carried out using internationally standardized methods, following a common protocol 50 . Information on demographics, socioeconomic status, and lifestyle factors was collected with standardized questionnaires. Alcohol consumption was categorized into groups of grams of alcohol per day: 0 (non-drinker), <25 and ≥25. Smoking status was categorized into groups of cigarettes per day: 0 (non-smoker), <20 and ≥20. Several parameters such as blood pressure, height, weight, waist circumference, and BMI were measured. The methods of measuring above parameters were referred to previous studies 51 .  Statistical analyses. The statistical analyses were performed with the statistical software package SPSS 22.0 (SPSS Inc., Chicago, Illinois). The quantitative variables were presented as mean ± standard deviation (serum TG levels were presented as medians and interquartile ranges for not a normal distribution). Allele frequency was determined via direct counting, and the Hardy-Weinberg equilibrium was verified with the standard goodness-of-fit test. The genotype distribution between the two groups was analyzed by the chi-square test. General characteristics between two ethnic groups were compared by the Student's unpaired t-test. The association between genotypes and serum lipid parameters was tested by covariance analysis (ANCOVA). Any SNPs associated with serum lipid profiles at the value of P < 0.025 (corresponding to P < 0.05 after adjusting for 2 independent tests by the Bonferroni correction) were considered statistically significant. Gender, age, BMI, blood pressure, alcohol consumption and cigarette smoking were adjusted for the statistical analysis. Haploview (Broad Institute of MIT and Harvard, USA, version 4.2) analyzed the haplotype frequencies and pair-wise LD among the detected SNPs. Unconditional logistic regression was used to assess the correlation between the risk of hyperlipidemia and genotypes. Multivariable linear regression analyses with stepwise modeling were used to determine the correlation between the genotypes (GG = 1, GC = 2, CC = 3) and several environmental factors with serum lipid levels in males and females of Han and Maonan populations. Two sides P value < 0.05 was considered statistically significant. The heart-map of inter-locus models was measured by R software (version 3.3.0) 58 .