Positive Association between APOA5 rs662799 Polymorphism and Coronary Heart Disease: A Case-Control Study and Meta-Analysis

Objective Apolipoprotein A5 (APOA5) is associated with plasma triglyceride (TG) levels, a risk factor for coronary heart disease (CHD). This study explored the association between CHD and the APOA5 rs662799 polymorphism. Methods We collected 1,521 samples (783 CHD patients and 738 controls) for this case-control study. Meta-analysis was performed using Review Manager Software and Stata Software. Results Significant differences were observed between CHD cases and controls at the level of both genotype (χ2 = 8.964, df = 2, P = 0.011) and allele (χ2 = 9.180, df = 1, P = 0.002, OR = 1.275, 95% CI = 1.089–1.492). A breakdown analysis by gender showed a significant association of APOA5 rs662799 with CHD in males (χ2 = 7.770, df = 1, P = 0.005; OR = 1.331, 95% CI = 1.088–1.628). An additional meta-analysis using 21378 cases and 28428 controls established that rs662799 is significantly associated with CHD (P < 0.00001). Conclusion Both our case-control study and meta-analysis confirm a significant association between APOA5 rs662799 and CHD. In addition, our results suggest a male-specific association between the APOA5 rs662799 polymorphism and CHD.


Introduction
PCR System 9700 Dual 384-Well Sample Block Module (Applied Biosystems, Foster City, CA, USA). PCR conditions included an initial denaturation of 95°C for 2 min, followed by 45 cycles of 95°C for 30 sec, 56°C for 30 s, 72°C for 1 min and then a final extension at 72°C for 5 min. After purification by SAP Reaction, we proceeded with primer extension. The primer extension protocol included an initial denaturation at 94°C for 30 s, followed by 40 cycles of amplification (including 94°C for 5 s, 52°C for 5 s, 80°C for 5 s), 5 cycles of amplification (5 s at 52°C, 5 s at 80°C), a final extension at 72°C for 3 min after which samples were held at 4°C. Single nucleotide polymorphism genotyping was performed using the Sequenom Mass-ARRAY iPLEX platform per the manufacturer's instructions [28]. The primer sequences were 5'-ACGTTGG ATGAGCATTTGGGCTTGCTCTCC-3' (first primer), 5'-ACGTTGGATGTCTGAGCCCCA GGAACTGGA-3' (second primer) and 5'-caGAACTGGAGCGAAAGT-3' (extended primer).

Publication retrieval and data extraction
The literatures were searched in the online databases including PubMed and Wanfang between Jan 2000 and Jul 2015. The keywords were "coronary heart disease", "coronary artery disease" or "myocardial infarction" combined with "APOA5" and "rs662799" or "-1131T>C". All of the case-control studies between APOA5 rs662799 and CHD were retrieved for the consideration of the current meta-analysis. All of the case-control studies between APOA5 (rs662799) and CHD were considered to be eligible for the current meta-analysis. We only included studies that presented data on allele or genotype frequencies for both cases and controls and displayed a genotype distribution meeting Hardy-Weinberg equilibrium (HWE) [29]. Information in the meta-analysis included the first author's name, publication year, country, ethnic group, number of alleles or genotypes and the total number of cases and controls. The details on the inclusion criteria included as follow:1) only the case-control studies on the association between rs662799 and CHD were included; 2) the eligible studies must contain the odds ratios (ORs) and 95% confidence intervals (CIs), or the genotype or allele information to calculate ORs and 95% Cis; 3) HWE should be met for the genotype distribution in the control group of the eligible studies if they have genotype information. We directly emailed the corresponding authors or called them (only for authors in China) for the missing information in their studies. There were 214 studies retrieved from the Wanfang and CNKI literature databases after searching for the keywords "coronary heart disease", "coronary artery disease" or "myocardial infarction" combined with "APOA5" and "rs662799" or "-1131T>C". After a series of selection procedures, we excluded 15 duplicate studies, 5 meta-analysis studies, 120 irrelevant studies, 28 studies on other diseases, and 7 studies without genotyping data (S1 File). In addition, we further downloaded the GWAS dataset from WTCCC research, then we imputed the information of rs662799 genotype by MaCH-Admix in WTCCC database, and added the data to the metaanalysis [30]. The remaining 40 case-control studies were qualified for our meta-analysis ( Fig  1) [11,13,17,20,22,.

Statistical analyses
The HWE test was performed using the Arlequin program (version 3.5), and P > 0.05 was considered to be in HWE [64]. Genotype and allele distribution was compared between cases and controls by CLUMP22 software using 10,000 Monte Carlo simulations [65]. The odds ratio (OR) with a 95% confidence interval (CI) were determined using an online program, (http:// faculty.vassar.edu/lowry/odds2x2.html) [23]. Meta-analysis was performed using the Review Manager software set to the fixed-effect or random-effect method (version 5.0, Cochrane Collaboration, Oxford, United Kingdom) [66]. Heterogeneity in the meta-analysis was assessed using the Q and I 2 tests. An I 2 > 50% indicated the existence of heterogeneity among the studies in the meta-analysis. Publication bias was shown by Begg's funnel plot analysis, which was generated with Stata software (version 11.0, Stata Corporation, College Station, TX, USA). P values < 0.05 were significant.

Results
No departure from HWE was observed for the APOA5 rs662799 polymorphism in cases (P = 0.220) or controls (P = 0.544). Genotypic and allelic comparisons between cases and controls are shown in Table 1. Our data show that rs622799 is associated with the risk of CHD (genotype: χ 2 = 8.964, df = 2, P = 0.011; allele: P = 0.002; OR = 1.275, 95% CI = 1.089-1.492). A further gender-stratified association shows that rs662799 is significantly associated with CHD

Meta-analysis
Searching the existing literature databases, we found 40 case-control studies, 30 more cases than were used in the most recently published meta-analysis in 2013 [23]. Therefore, we performed an updated meta-analysis to investigate the link between rs662799 and CHD. Information from these 40 eligible studies and our case-control study are shown in Table 3. Among the 40 eligible studies in the current meta-analysis, 7 studies only had allelic information. Therefore, allele-based model was applied in the meta-analysis. For the meta-analysis with moderate heterogeneity (I 2 < 50%), we selected a fixed-effect model for the meta-analysis, otherwise, the random-effect model was used for the meta-analysis with great heterogeneity (I 2 > = 50%). The current meta-analysis has great heterogeneity (I 2 = 70%), therefore random-effect model was used. As shown in Fig 2, subgroup meta-analysis by major ethnic groups also indicates a significant association between APOA5 rs662799 and CHD in Asians (P = 0.01, I 2 = 66%), Chinese (P < 0.000001, I 2 = 67%) and Caucasians (P = 0.008, I 2 = 60%). The meta-analyses show no publication bias by Begg's funnel plot analysis (Fig 3). Furthermore, sensitivity analysis suggests that the conclusion is not biased by any individual study (Fig 4).

Discussion
Our results show that the rs662799 polymorphism in the APOA5 gene is significantly associated with CHD in Han Chinese (P = 0.011). The minor G allele of APOA5 rs662799 may increase the risk of CHD by 27.5% (P = 0.002, OR = 1.275, 95% CI = 1.089-1.492). Consistent with previous reports, the rs662799-G allele is associated with higher leaves of TG in both CHD patients and controls [13,67]. A power calculation for APOA5 rs662799 indicates that our study has 85.9% power to detect significance in the association test.
Environmental factors, such as gender and age, are important factors of CHD. The prevalence of CHD in females was different from males [2,68,69]. Evidence has shown that patients older than 65 years have a higher cardiovascular morbidity and mortality [70,71]. In the current meta-analysis, we were unable to perform the subgroup meta-analysis by the age or gender due to a paucity of related information in the involved studies. Gender and age are independence risk of CHD [4,[72][73][74]. Epidemiologic evidence suggests that the risk of morbidity and mortality are higher in male CHD patients than in females [75]. Our data show a strong association between APOA5 rs662799 and CHD in the male group, providing a novel molecular explanation for the gender disparity observed in CHD. In addition, we showed a statistically significant difference between rs662799 and CHD in the subgroup aged from 55-65, although the underlying mechanism will require additional studies. The frequency of the APOA5 rs662799 polymorphism varies greatly among different populations. The rs662799-G allele frequency is 26.7% in Chinese populations, similar to that in Japanese populations (29.1%). However, the Chinese frequency is much higher than that in  European populations (1.7%). Nevertheless, accumulating evidence indicates a strong association between APOA5 rs662799 and CHD among different populations. In addition to APOA5 rs662799, there are associations between other APOA5/A4/C3/A1 polymorphisms and CHD, which include APOA5 rs3135506 and APOA/A4/C3/A1 cluster haplotypes [76]. Further functional analysis is needed to discriminate the relationship among these polymorphisms.
There were other seven APOA5 polymorphisms involved in the genetic studies (S2 Table). However, rs3135506 (n = 7) [43,45,49,53,56,76,77] and -12238T/C (n = 1) [78] were tested for the association of CHD. Thus, we only included rs662799 in the current meta-analysis. Among the published GWAS related to the current meta-analysis, we didn't find any direct information that could be applied in the current meta-analysis [12,67,79]. We further and added the WTCCC data to the meta-analysis. Please see the following figure for the updates (Fig 2). The current meta-analysis includes 40 studies comprised of 21378 cases and 28428 controls from 10 ethnic populations. Our meta-analysis contains at least 26 case studies and 3 ethnic populations more than were included in the last five meta-analyses published [23,[80][81][82][83]. All of the meta-analyses indicate that the APOA5 rs662799 polymorphisms associated with CHD in the Chinese population, although many of the studies did not include a subgroup analysis stratified by ethnicity.
Despite the merits of our meta-analysis, there are limitations that must be considered. Our meta-analysis only includes studies from Asian and Caucasian populations. Therefore, it might not be an accurate representation of other ethnicities, such as African populations. Publication and language bias might exist in the case control studies [84]. The current meta-analysis was involved with 10 Caucasian and 30 Asian studies. Among the Asian studies, there were 24 Chinese studies (7 in English and 17 in Chinese). A further check for the minor allele frequency report in the HapMap International Project, we found the MAF in Europeans was 1.7% which was much less than 26.7% in Chinese and 29.1% in Japanese. However, subgroup meta-analyses by ethnicity found significant association of APOA5 rs662799 and CHD in both Europeans and Asians. There may also be a selection bias in our meta-analysis, which only included studies published in English or Chinese. Finally, standards for diagnosis may vary due to differences in the inclusion of CHD cases and non-CHD controls.
In summary, our case-control and meta-analysis demonstrates that the frequency of the APOA5 rs662799-G allele is significantly increased in CHD cases compared with controls. Furthermore, APOA5 rs662799 interacts with both gender and age in the association with CHD.  Table. Other seven APOA5 polymorphisms involved in the genetic studies. (DOC)