Association between the DOCK7, PCSK9 and GALNT2 Gene Polymorphisms and Serum Lipid levels

This study was to determine the association between several single nucleotide polymorphisms (SNPs) in the dedicator of cytokinesis 7 (DOCK7), proprotein convertase subtilisin/kexin type 9 (PCSK9) and polypeptide N-acetylgalactosaminyltransferase 2 (GALNT2) and serum lipid levels. Genotyping of 9 SNPs was performed in 881 Jing subjects and 988 Han participants. Allele and genotype frequencies of the detected SNPs were different between the two populations. Several SNPs were associated with triglyceride (TG, rs10889332, rs615563, rs7552841, rs1997947, rs2760537, rs4846913 and rs11122316), high-density lipoprotein (HDL) cholesterol (rs1997947), low-density lipoprotein (LDL) cholesterol (rs1168013 and rs7552841), apolipoprotein (Apo) A1 (rs1997947), ApoB (rs10889332 and rs7552841), and ApoA1/ApoB ratio (rs7552841) in Jing minority; and with TG (rs10889332, rs615563, rs7552841, rs11206517, rs1997947, rs4846913 and rs11122316), HDL cholesterol (rs11206517 and rs4846913), LDL cholesterol (rs1168013), ApoA1 (rs11206517 and rs4846913), ApoB (rs7552841), and ApoA1/ApoB ratio (rs4846913) in Han nationality. Strong linkage disequilibria were noted among the SNPs. The commonest haplotype was G-C-G-C-T-G-C-C-G (>10%). The frequencies of C-C-G-C-T-G-T-C-G, G-C-A-C-T-G-C-C-G, G-C-G-C-T-A-C-C-A, G-C-G-C-T-G-C-C-A, G-C-G-C-T-G-T-C-A haplotypes were different between the two populations. Haplotypes could explain much more serum lipid variation than any single SNP alone especially for TG. Differences in lipid profiles between the two populations might partially attribute to these SNPs and their haplotypes.

a role in axon formation and neuronal polarization. The encoded protein displays GEF activity toward RAC1 and RAC3 Rho small GTPases but not toward CDC42. Several transcript variants encoding different isoforms have been found for this gene. The proprotein convertase subtilisin/kexin type 9 (PCSK9; Gene ID: 255738; MIM: 607786) gene, also known as FH3, PC9, NARC1, LDLCQ1, NARC-1 and HCHOLA3, is located on 1p32. 3 and this gene encodes a member of the subtilisin-like proprotein convertase family, which includes proteases that process protein and peptide precursors trafficking through regulated or constitutive branches of the secretory pathway. The encoded protein undergoes an autocatalytic processing event with its prosegment in the ER and is constitutively secreted as an inactive protease into the extracellular matrix and trans-Golgi network. It is expressed in liver, intestine and kidney tissues and escorts specific receptors for lysosomal degradation. It plays a role in cholesterol and fatty acid metabolism. Mutations in this gene have been associated with autosomal dominant familial hypercholesterolemia. Alternative splicing can result in multiple transcript variants. The polypeptide N-acetylgalactosaminyltransferase 2 (GALNT2; Gene ID: 2590; MIM: 602274) gene, formerly known as GalNAc-T2, is located on chromosome 1q41-q42 and encodes a member of the glycosyltransferase 2 protein family. Members of this family initiate mucin-type O-glycoslation of peptides in the Golgi apparatus. The encoded protein may be involved in O-linked glycosylation of the immunoglobulin A1 hinge region. This gene may influence TG levels, and may be involved type 2 diabetes, as well as several types of cancer. Alternative splicing can also result in multiple transcript variants (http://www.ncbi.nlm.nih.gov/gene/).
Human genetic studies of lipid levels can identify targets for new therapies for cholesterol management and prevention of heart disease especially monoclonal anti-PCSK9 antibodies are already on the market to significantly reduced levels of LDL cholesterol when added to statin therapy administered at the maximum tolerated dose 14 . For comparison with the nonsynonymous single nucleotide polymorphisms (SNPs) in known drug therapies genes, we scored point mutations at synonymous point mutations in housekeeping genes or genes of unknown function on the approximate locations of the chromosome 1. The exact positions of the PCSK9 SNPs were located in the similar position area of the DOCK7 and GLANT2 SNPs (http://hapmap.ncbi.nlm.nih.gov/). Several genetic variants in the DOCK7, PCSK9 and GLANT2 have been associated with serum lipid parameters, especially with TG in Western populations 15 , e.g. the SNPs of DOCK7 rs1167998, rs10889353 16 , PCSK9 rs11591147 17 and GALNT2 rs4846914 13 were associated with TG levels in European and PCSK9 rs505151 18 and GALNT2 rs2144300 and rs4846914 19 in the Asian populations. However, the association of the DOCK7 (rs1168013 and rs10889332), PCSK9 (rs615563, rs7552841 and rs1126517) and GALNT2 (rs1997947, rs2760537, rs4846913 and rs11122316) SNPs and serum lipid levels has not been previously reported. Since ancient times China is a multi-ethnic country. Among 56 nationalities in China, the Han nationality is the biggest one. Jing is one of the smallest population of ethnic minorities in southern China with a population of 22,517 (in 2000 the fifth national census statistics of China), China's only a coastal fishery ethnic minority, and China's only national ocean at the same time 20 . Jing populations live in Dongxing city, Guangxi Zhuang Autonomous Region. Diet to rice is given priority to, fresh fish and shrimp more, like to use fish sauce to taste. The history of Jing ethnic minority shows Jing nationality is a relatively conservative and isolated minority, and preserves their custom of intra-ethnic marriage 21 . Thus, their genetic background may be less heterogeneous within the population. Little is known about the association of SNPs and lipid phenotypes in the Jing population. Therefore, this research was undertaken to detect the association of the DOCK7 rs1168013, DOCK7 rs10889332, PCSK9 rs615563, PCSK9 rs7552841, PCSK9 rs11206517, GALNT2 rs1997947, GALNT2 rs2760537, GALNT2 rs4846913, and GALNT2 rs11122316 SNPs and lipid profiles in the two ethnic groups.
Haplotypes and lipid profiles. The correlation of the haplotypes and lipid profiles is shown in Table 6. Rare Hap (frequency < 3%) in both Jing and Han populations has been dropped. The carriers of C-C-G-C-T-G-T-C-G haplotype had lower TG and higher HDL cholesterol levels in Jing plus Han populations and lower TG levels in Jing population than the non-carriers of C-C-G-C-T-G-T-C-G haplotype (P < 0.05). There were no differences in lipid parameters between the carriers and non-carriers of C-C-G-C-T-G-T-C-G haplotype in the Han population. Haplotype G-C-A-C-T-G-C-C-G carriers had lower serum TG in the Han populations than the haplotype non-carriers (P < 0.05). Haplotype G-C-G-C-T-A-C-C-A carriers had higher serum TG and lower ApoA1 levels in Jing plus Han population, and higher serum TG and lower HDL cholesterol and ApoA1 in Jing population than the haplotype G-C-G-C-T-A-C-C-A non-carriers (P < 0.05 for each). Haplotype G-C-G-C-T-G-C-C-A carriers had lower TC, TG, LDL cholesterol and ApoB levels in Jing plus Han population, lower TC, TG and ApoB in Jing ethnic minority and lower TG levels than the haplotype G-C-G-C-T-A-C-C-A non-carriers (P < 0.05 for all).
Correlation between lipid parameters and alleles or genotypes. Table 7 depicts the direction and magnitude of associations between lipid parameters and alleles or genotypes of the 9 SNPs in the Jing and Han populations. Adjusting for age, sex, BMI, smoking status, alcohol use, and exercise, logistic regression analysis showed that several the examined SNPs were significant correlated with lipid parameters.

Discussion
In the present study, we showed for the first time the association of the DOCK7 (rs1168013 and rs10889332), PCSK9 (rs615563, rs7552841 and rs1126517) and GALNT2 (rs1997947, rs2760537, rs4846913 and rs11122316) SNPs and some serum lipid parameters; the LD status and the haplotype frequencies of the detected SNPs.
In addition, we also successfully replicated the association of DOCK7 rs10889332, PCSK9 rs615563, PCSK9 rs7552841, GALNT2 rs1997947, GALNT2 rs2760537, GALNT2 rs4846913 and GALNT2 rs11122316 SNPs with the levels of serum TG in the Jing ethnic minority; and DOCK7 rs10889332, PCSK9 rs615563, PCSK9 rs7552841, PCSK9 rs11206517, GALNT2 rs1997947, GALNT2 rs4846913 and GALNT2 rs11122316 with serum TG levels in the Han nationality. The SNPs of rs636523 and rs12130333 22 near the DOCK7/ ANGPTL3, PCSK9 rs505151 23 and GALNT2 rs4846914 24,25 have been associated with TG in some previous studies, but the association of the 9 SNPs and other serum lipid parameters has not been reported previously. The genotype and allele frequencies of several SNPs in this study were also not reported previously in different racial/ethnic groups. In the present study, we revealed that the genotypic and allelic frequencies of the DOCK7 rs1168013, DOCK7 rs10889332, PCSK9 rs615563, PCSK9 rs7552841, PCSK9 rs1126517, GALNT2 rs1997947, GALNT2 rs2760537, GALNT2 rs4846913 and GALNT2 rs11122316 SNPs were different between the two ethnic groups. All of the detected SNPs were in the Hardy-Weinberg equilibrium except DOCK7 rs10889332. The minor allele or rare homozygote genotype frequencies of the 9 SNPs in the Han nationality were in close proximity to those of CHB from the international haplotype map (HapMap;http://hapmap.ncbi.nlm.nih.gov/cgi-perl/gbrowse/hapmap24_B36/) data. The minor allele or rare homozygote genotype frequencies of the 9 detected SNPs were also lower in European ancestries than in Asian nationalities from the data. These results suggest that the prevalence of the minor allele or rare homozygote genotype frequencies of the 9 SNPs may have a racial/ethnic-specificity.
In the present reasearch, our findings also showed that there may be a racial/ethnic specific association of the 9 SNPs and lipid parameters. The association of other SNPs near DOCK7, PCSK9 and GALNT2 and lipid profiles has been reported previously. Through fine-mapping, previous study discovered the SNP with significant associations, with consistent effect on TG levels across ancestral groups: rs636523 near DOCK7/ANGPTL3. African LD patterns did not assist in narrowing association signals 22 . PCSK9 (TG, HDL cholesterol, ApoB and ApoA1/ ApoB) was shown interactions with overweight/obesity to influence serum lipid levels 23 . Our team reported that the correlations of both GALNT2 rs2144300 and GALNT2 rs4846914 SNPs and lipid parameters were different between the two ethnic groups 19 . However, Several GWASs and candidate gene researches failed to find the association between the GALNT2 polymorphisms and lipid parameters [26][27][28] . There was no any effect of the GALNT2 rs4846914 on the levels of serum TC or TG reported by Polgár et al. previously 26 . In Whitehall II, there was a significant correlation of the GALNT2 variants and serum lipoprotein (a) levels. Whereas any of these findings did not confirmed in the previously meta-analysis of six studies 28 . It could be due to the effects of these SNPs were modest on serum lipid concentrations and/or lower statistical power to determine the correlation was present 26,29 . In addition, gene-environmental and environmental-environmental factors on lipid parameters remain to be interpreted.
Many GWASs have reported that the association of other variants near DOCK7, PCSK9 and GALNT2 and serum lipid levels is still controversial. Pleiotropic effects on the lipid profile, the potential correspondence was detected for ANGPTL3 30 and DOCK7 31 being highly associated with cholesterol and LDL cholesterol levels. Loss-of-function mutations in the ANGPTL3 were associated with decreased levels of LDL cholesterol, HDL  cholesterol and TG 32 . The associations observed for the DOCK7 locus, which is involved in neurogenesis, myelination and axon formation 33 but not in lipid metabolism probably reflect the co-localization of this gene with ANGPTL3. As expected, rare variants that contribute to population differences tend to be population specific, exemplified by multiple African-specific variants in PCSK9 associated with LDL cholesterol 34 . The SNPs in intron 1 of a GalNac transferase (GALNT2) were identified as a novel lipid-associated region from GWAS and subsequent knock-down and overexpression of this gene in mouse liver clearly demonstrated that GALNT2 can influence HDL cholesterol levels 35 .
The cause of the contradictions in correlation of the detected SNPs with lipid parameters among the different population is not completely understood. This could be because of the differences in genetic background in some degree. Compared to the Han nationality, the Jing ethnic minority had higher the value of weight, BMI, waist circumference, the serum TC and TG levels and the lower percentage of participants who consumed alcohol and the ratio of ApoA1 to ApoB. Among 56 nationalities in China, Han nationality is the largest one. Jing ethnic minority was less population nationality with the population of 22517 according to the China's fifth national census in 2000. Approximately 90% of the Jing people live in the three islands of Wanwei, Wutou and Shanxin in the Dongxing city, Guangxi, China. About 1511, their ancestors emigrated from Vietnam to China and first settled on the aforementioned three islands. Therefore, some hereditary background and alleles/genotypes of lipid metabolism-related genes in Jing ethnic minority might be somewhat different from those in Han nationality. Another reason could be because of the ethnic difference in their LD pattern. In this research, we detected that the frequencies of the C- haplotypes were significantly different between the Jing and Han populations. The haplotypes with nine SNPs could explain much more serum lipid variation than any single SNP alone, especially for TG. Therefore, ethnic differences in the LD pattern could partially explain the discrepancy in the correlation of the detected SNPs with lipid parameters among diverse nationalities.
Several environmental factors independently such as hypertension, obesity, physical activity, dietary patterns and lifestyle are related with lipid parameters strongly [36][37][38][39][40][41][42] . There was association of gender, age, BMI, cigarette smoking, alcohol consumption, blood pressure and lipid levels in both Jing and Han populations. These detected data determined some environmental factors play an important role in determining lipid parameters. For approximately half a century it has been acknowledged that diets of high-fat particularly contain the large quantities of saturated fatty acids raise predispose individuals to hyperlipidemia and        CVD 43 . In the current study, we found that the % of participants who consumed alcohol were lower in Jing than in Han nationality. Although effects of the alcohol consumption on lipid parameters appear to vary by types of specific patient or the alcohol intake patterns, and perhaps by sex and population, the subject research has been the focus of so much current studies [44][45][46] .
GWASs have identified many loci that will harbor genes relevant to the biology of lipid levels. The results show that significant associations can be identified by studying a relatively small number of subjects with extreme values of a quantitative lipoprotein trait. Re-sequencing genes at GWAS loci may reveal new rare loss-of-function mutations, creating potential new therapeutic targets for decreasing the prevalence of heart disease. Until recently, most genome-wide efforts have used genotyping arrays and imputation to assay most of the common variation across the genome. Recent technological advances have enabled whole genome sequencing approaches, which hold the promise of discovery of novel rare variants with large effects on lipid levels and heart disease risk.
There are several potential limitations in the present study. Firstly, the sample size is a bit small as compared with many previous GWASs. Hence, further studies with larger sample sizes are needed to confirm our results. Secondly, interactions of gene-environmental or environmental-environmental factors on serum lipid traits remain to be detected. Thirdly, there are no independent haplotypes of each gene considered instead. Phasing of the 9 SNPs is so far away on chromosome 1 (~40MB). There will be many recombination events that will be missed due to the lack of information across the region. Finally, although we have detected the effects of 9 SNPs in the PCSK9, DOCK7 and GALNT2 on serum lipid levels in this study, there are still many lipid-related SNPs and the interactions of SNP-SNP and/or SNP-environmental factors. What's more, the relevance of this finding has to be defined in further high caliber of studies including incorporating the genetic information of the DOCK7, PCSK9 and GALNT2 SNPs and their haplotypes and in vitro functional studies to confirm the impact of a variant on a molecular level.

Materials and methods
Subjects and research design. For the current study, 881 unrelated subjects (456 males, 51.76% and 425 females, 48.24%) of Jing and 988 (536 men, 54.25% and 452 women, 45.75%) unrelated individuals of Han from Dongxing city, Guangxi Zhuang Autonomous Region, China were selected randomly from our randomized, stratified samples. The subjects' age ranged from 15 to 80 years and with the average age of 56.69 ± 13.39 years in Jing and 56.18 ± 12.85 years in Han. All participants were healthy with no disease history of atherosclerosis, CVD, diabetes, thyroid and/or kidney. When blood samples were taken, none of them used lipid-modulating therapy such as fibrates or statins. The study design was approved by the Ethics Committee of the First Affiliated Hospital, Guangxi Medical University. Written informed consent was obtained from all participants. All experiments were performed in accordance with relevant guidelines and regulations [47][48][49][50] .

Data collection
Epidemiological investigation and measurements of biochemical markers. Participants participated in baseline examination conducted in the study center by trained staff following standardized protocols, which included anthropometric, blood pressure measurements, height, weight (without shoes) and waist circumference parameters (in cm was measured at the midpoint between the lower ribs and the iliac crest), a blood sample collection as well as a personal interview on medical history, a sociodemographic, socioeconomic status and lifestyle questionnaire and a self-administered food frequency questionnaire; BMI was calculated   Table 5. Lipid profiles according to genotypes for the two ethnic groups. HDL, high density lipoprotein; LDL, low density lipoprotein; The P-value calculated by ANCOVA, using general linear models, and adjusted for age, sex, BMI, smoking status, alcohol use, glucose and hypertension, and less than 0.005 was considered statistically significant after adjusting by Bonferroni correction. n = sample size.
Scientific  2) analyzed the haplotype frequencies and pair-wise LD among the detected SNPs. The correlation of genotypes and lipid profiles was calculated by ANCOVA. Any SNPs associated with the lipid profiles at the value of P < 0.005 (corresponding to P < 0.05 after adjusting for 9 independent tests by the Bonferroni correction) were considered statistically significant. Unconditional logistic regression was used to assess the assocation between lipid parameters and genotypes (common homozygote genotype = 1, heterozygote genotype = 2, rare homozygote genotype = 3) or alleles (the minor allele non-carrier = 1, the minor allele carrier = 2). Gender, age, BMI, alcohol consumption, cigarette smoking and hypertension were adjusted for statistical analysis. Two-sided P value of less than 0.05 was considered statistically significant.