Identification of Susceptibility Variants in ADIPOR1 Gene Associated with Type 2 Diabetes, Coronary Artery Disease and the Comorbidity of Type 2 Diabetes and Coronary Artery Disease

Objective Adiponectin receptor 1 (encoded by ADIPOR1) is one of the major adiponectin receptors, and plays an important role in glucose and lipid metabolism. However, few studies have reported simultaneous associations between ADIPOR1 variants and type 2 diabetes (T2D), coronary artery disease (CAD) and T2D with CAD. Based on the “common soil” hypothesis, we investigated whether ADIPOR1 polymorphisms contributed to the etiology of T2D, CAD, or T2D with CAD in a Northern Han Chinese population. Methods Our multi-disease comparison study enrolled 657 subjects, including 165 with T2D, 173 with CAD, 174 with both T2D and CAD (T2D+CAD), and 145 local healthy controls. Six ADIPOR1 single nucleotide polymorphisms (SNPs) were genotyped and their association with disease risk was analyzed. Results Multi-case-control comparison identified two ADIPOR1 variants: rs3737884-G, which was simultaneously associated with an increased risk of T2D, CAD, and T2D+CAD (P-value range, 9.80×10−5−6.30×10−4; odds ratio (OR) range: 1.96–2.42) and 16850797-C, which was separately associated with T2D and T2D+CAD (P-value range: 0.007–0.014; OR range: 1.71–1.77). The risk genotypes of both rs3737884 and 16850797 were consistently associated with common metabolic phenotypes in all three diseases (P-value range: 4.81×10−42−0.001). We observed an increase in the genetic dose-dependent cumulative risk with increasing risk allele numbers in T2D, CAD and T2D+CAD (P trend from 1.35×10−5−0.002). Conclusions Our results suggest that ADIPOR1 risk polymorphisms are a strong candidate for the “common soil” hypothesis and could partially contribute to disease susceptibility to T2D, CAD, and T2D with CAD in the Northern Han Chinese population.


Introduction
Coronary artery disease (CAD), type 2 diabetes (T2D), and T2D with CAD are multifactorial diseases in which hereditary and environmental factors both contribute to their etiology. These diseases may have a common pathogenesis based on the ''common soil'' hypothesis in which diabetes and cardiovascular disease share common antecedents [1]. Indeed, CAD, one of the main causes of death worldwide [2], and T2D together lead to the development of T2D with CAD.
Adiponectin receptor 1, encoded by the ADIPOR1 gene, is a major adiponectin receptor that mediates the glucose and lipid metabolism-related effects of adiponectin on target cells. Research based on animal models has shown that ADIPOR1 overexpression can augment the biological effects of adiponectin [3], whereas ADIPOR1 knockout leads to increased insulin resistance (IR) and endogenous glucose production [4], suggesting a correlation between ADIPOR1 expression and adiponectin activity [5]. Moreover, Wang et al. showed that down-regulated ADIPOR1 signaling was the underlying mechanism for increased foam cell formation and accelerated cardiovascular disease in diabetic subjects [6].
Although several association studies reported that ADIPOR1 variants were risk factors for IR [7,8] or T2D [9][10][11], few studies investigated the relationship between ADIPOR1 polymorphisms and CAD [12] or T2D with CAD [13]. In particular, there are limited reports about ADIPOR1 variant simultaneous associations with T2D, CAD and T2D with CAD.
Based on the above-mentioned ''common soil'' hypothesis, we hypothesized that the etiology of T2D, CAD, and T2D with CAD could at least partially be associated with ADIPOR1 polymorphisms, which may affect the interaction between receptor and ligand and thus play crucial roles in the development of genetic variants associated with these three diseases. We therefore conducted a multi-case-control association study to investigate the relationship between common ADIPOR1 variants and the three diseases status in the Northern Han Chinese population.

Ethics Statement
This study complies with the Declaration of Helsinki, and the local ethics committees of the two participating hospitals (Beijing Anzhen Hospital, Capital Medical University, Beijing, China; Beijing Hospital & Beijing Institute of Geriatrics, Chinese Ministry of Health, Beijing, China) approved the research protocol. Written informed consent was obtained from each participant.

Study Populations
This population-based, multi-case-control study was conducted on subjects who were permanent residents of the Beijing area, China, of self-identified Han ethnic origin. We enrolled a total of 657 individuals: 165 patients with T2D (T2D group), 173 with CAD (CAD group), 174 with both T2D and CAD (T2D+CAD group) and 145 healthy controls (Control group). Patients from the CAD and T2D+CAD groups were hospitalized at Beijing Anzhen Hospital between March 2007 and December 2009, while participants of the T2D and Control groups were recruited from Beijing Hospital between June and December 2008. Participant characteristics are given in Table 1. T2D was diagnosed according to World Health Organization criteria [14], while classification of CAD patients was based on previous studies [15]. Patients of the T2D+CAD group met both the above-mentioned inclusion criteria. The inclusion criteria of the Control group were as follows: no ascertained diabetes, CAD, or T2D with CAD in first-degree relatives; no dyslipidemia, abnormal glucose tolerance, or high blood pressure; no potential CAD or myocardial infarction as determined from medical records or electrocardiographic tests. Subjects from all groups except the Control group underwent standard coronary angiography.
Individuals were excluded from the current study if: (1) they were #18 years old; (2) they had type 1 diabetes, diabetic ketoacidosis, hyperosmolar nonketotic diabetic coma, were rele-vantly autoantibody-positive, or needed insulin injections during the first year; (3) they had heart failure, myocardiopathy, or congenital heart disease; (4) they had autoimmune diseases (such as rheumatic diseases), cancer, infection, severe liver or renal diseases, were pregnant, or currently using glucocorticoid.

Clinical Measurements and Biochemical Analyses
All study subjects were examined in the morning after an overnight fast to take anthropometric measurements and to collect blood samples for biochemical measurements and DNA extraction. Systolic blood pressure (SBP), diastolic blood pressure (DBP), height, weight, body mass index (BMI, calculated as weight in kilograms divided by the square of height in meters), plasma levels of total cholesterol (TC), triglycerides (TG), fasting plasma glucose (FBG), plasma high-density lipoprotein cholesterol (HDL-C) and plasma low-density lipoprotein cholesterol (LDL-C) were measured or calculated as previously described [16].

Genotyping
Blood samples of participants were collected by standard venipuncture into evacuated vacuum tubes with ethylene diaminetetraacetic acid (EDTA). Genomic DNA was extracted from whole blood samples using standard DNA isolation methods [17]. The concentration and purity of the extracted DNA was determined by NanoDrop 2000c (Thermo Fisher, USA).
PCR was performed using a Bio-Rad C1000TM (Bio-Rad, USA). The 10 mL RCR reaction mixture for PCR-RFLP consisted of 0.1 mL forward and reverse primers (10 pmoL/mL), 1 mL DNA template, 1 mL 106PCR Buffer, 0.1 mL Taq DNA polymerase (5 U/mL), 0.2 mL dNTP (10 mmoL/L), and deionized water to the total volume. PCR conditions were as follows: initial denaturation at 95uC for 5 min, followed by 35 cycles of denaturation at 95uC for 30 s, annealing for 30 s (Table S1), and extension at 72uC for 30 s, followed by 72uC for 7 min. Subsequently, PCR products were digested for 4 h at 37uC in a 10 mL mixture of 3 mL PCR product, 1 mL 106PCR Buffer, 0.3 mL restriction enzyme (Table S1) and 5.7 mL deionized water, followed by detection on 8% ethidium bromide-stained polyacrylamide gel electrophoresis ( Figure S1).
PCR reaction mixture for PCR-HRM was carried out in a total volume of 10 mL, including 0.18 ml 16LC Green plus (Idaho, USA), 0.02 ml forward and 0.18 mL reverse primers (10 pmoL/ml), 1 ml DNA template, 1 mL 106PCR Buffer, 0.1 mL Taq DNA polymerase (5 U/mL), 0.2 mL dNTP (10 mmoL/L), and deionized water to the total volume. PCR conditions were similar to those of PCR-RFLP but added two cycles at 95uC for 30 s, at 25uC for 2 min, at 94uC for 30 s, and at 24uC for 4 min. PCR products were transferred into HRM-specific 96-well plates, genotyped automatically, and verified manually using a LightScanner TMHR-I 96 (Idaho, USA) ( Figure S1).
To check for errors in genotyping, 10% of each SNP samples were performed in duplicate and three samples of each genotype were randomly selected to be directly sequenced. The 50 mL reaction mixture included 0.5 mL of each primer (10 pmoL/mL), 5 mL DNA template, 5 mL 106PCR Buffer, 0.5 mL Taq DNA polymerase (5 U/mL), 1 mL dNTP (10 mmoL/L), and deionized water to the total volume. PCR conditions involved initial denaturation at 95uC for 5 min, followed by 35 cycles at 95uC for 30 s, annealing for 30 s (Table S1), extension at 72uC for 45 s, followed by 72uC for 7 min. PCR products were subjected to electrophoresis on 8% polyacrylamide gels, visualized using the gel imaging system (Gel Doc2000, Bio-Rad, USA), and then sequenced by Beijing Tianyi Huiyuan Biosience & Technology Inc (Beijing, China). No discrepancies were observed. To minimize misclassification bias, genotyping was performed blindly to all other data.

Statistical Analysis
Continuous variables were presented as median values and interquartile ranges or means 6 standard deviation, and categorical variables were presented as percentages, depending on the distribution of the variables. Differences in demographic and clinical characteristics between groups were compared using parametric (Student's t-test or one-way ANOVA test for normally distributed variables) or nonparametric (Mann-Whitney U test or Kruskal-Wallis test for non-normally distributed variables) methods for continuous variables, and Pearson's x 2 analysis or unconditional logistic regression for categorical variables. Individuals with missing data for a particular analysis were removed from the analysis. Tests for Hardy-Weinberg equilibrium (HWE) were performed separately for each SNP in control subjects using the online computer platform SHEsis (http://analysis.bio-x.cn/ myAnalysis.php). Pairwise LD was performed by Haploview V 4.1 (Broad Institute, Cambridge, MA) ( Figure 1). The risk-allele was designated 1 and the non-risk-allele 0. Genotypes were coded 0, 1 and 2, to represent the number of risk alleles carried by the subject, i.e., 2 represents carriers with two risk-alleles. Genotype frequencies were compared between cases and control subjects under the additive genetic model (comparing risk homozygous vs. heterozygous vs. wild homozygous carriers of variant) by unconditional logistic regression analysis adjusted for gender, age and BMI. The dominance genetic model (comparing risk homozygous + heterozygous vs. wild homozygous carriers of variant) and recessive genetic model (comparing risk homozygous vs. heterozygous wild + homozygous carriers of variant) analyses were also adopted. The cumulative effects of significant SNPs from single SNP analyses were calculated by counting the number of carriers with risk alleles. Odds ratios (OR) and 95% confidence intervals (CI) were used to estimate the strength of association between variables. The P-value for trend (P trend ) was obtained by performing a x 2 test for linear trend in EpiInfo version 6. SPSS statistical software package version 19.0 was used for statistical analysis. P-values were based on a two-sided test with the significance level set at P#0.05. Correction for multiple testing was performed by Bonferroni correction (http://www. quantitativeskills.com/sisa/calculations/bonfer.htm). Under this method, three independent hypotheses about a set of data were tested with a significance level set at P#0.017 (0.05/3). Table 1 shows the demographic and clinical characteristics of participants in the different groups, including median ages (58-65 years), the percentage of males (57.3-87.3%), and BMI means (22.4-26.5 kg/m 2 ), etc. Most of these data items were statistically different (P,0.05). Details of between-group comparisons were shown in Table S2. All SNPs examined in this study had three types of genotypes. PCR-RFLP ( Figure S1) and PCR-HRM ( Figure S2) results were consistent with those of sequencing analysis ( Figure S3). Allelic frequencies are summarized in Table  S3, and minor allele frequencies ranged from 0.11-0.40. Overall, each SNP conformed to HWE (P.0.05) in healthy controls, similar to that reported for HCB and Southern Han Chinese (CHS) in HapMap or Ensembl databases, but different from those reported for Utah residents with Northern and Western European ancestries from the CEPH collection (CEU) ( Table S4). The Haploview program revealed that these SNPs were in a weak pairwise LD with each other (Figure 1), so we included all of them in further analyses.
Genetic analysis for the association of these two ADIPOR1 SNPs with T2D, CAD, and T2D+CAD using dominant and recessive models achieved similar results to those seen for the additive model (Table S5).

Stratification Analysis of Associated SNPs by T2D or CAD Status
To assess whether there was an overlapping genetic effect of the SNPs on CAD or T2D risk, we performed a case-case comparative analysis stratified by T2D or CAD status (Table 3). Genotype frequencies of rs3737884 showed no significant differences between T2D and T2D+CAD groups or between CAD and T2D+CAD groups (P.0.05). Additionally, the genotype frequencies of rs16850797 did not differ significantly between T2D and T2D+CAD groups (P.0.05). However, the genotype frequencies of rs16850797 reached statistical significance in a comparison of CAD patients with and without T2D (P adjusted = 0.024, OR (95%CI) = 1.49 (1.05-2.10)) ( Table 3).

Stratification Analysis of Associated SNPs by Clinical Phenotypes
To further explore the possible effect of risk ADIPOR1 genotypes on clinical phenotypes, we selected and divided the subjects who carried risk genotypes into disease (T2D, CAD, and T2D+CAD) and healthy control subgroups. We defined the risk genotypes according to the dominance genetic model analysis in Table S5 and analyzed the association between clinical phenotypes and risk ADIPOR1 genotypes, including rs3737884 (GG+AG) and rs16850797 (CC+CG).
The multi-case-control comparison found significantly higher BMI, SBP, DBP, FBG and lower HDL-C levels in cases with rs3737884 (GG+AG) genotypes among T2D (P-values from 0.001 to 4.81610 242 ), CAD (P-values from 0.002 to 6.65610 239 ) and T2D+CAD (P-values from 3.4610 213 to 9.78610 238 ) groups than in the Control group. These same phenotypes were also more significant in patients with rs16850797-specific genotypes (CC+ CG) in T2D (P-values from 0.02 to 1.57610 212 ) and T2D+CAD (P-values from 4.63610 210 to 4.94610 233 ) groups compared with the Control group. (Table S6).

Cumulative Effect of Associated SNPs
Because haploview analysis did not reveal rs3737884 and rs16850797 to be in LD (Figure 1), we adopted a cumulative effect analysis to evaluate the genetic dose effect of risk alleles on T2D, CAD and T2D+CAD to clarify locus-locus interaction between risk alleles. As shown in Table 4, we observed an increase of the genetic dose-dependent cumulative risk with increased risk allele numbers. Overall, carriers with more than three risk alleles had a 2-3 fold risk for T2D (P = 1.64610 24 , OR (95%CI) = 2.93(1.67-5.17)), CAD (P = 0.002, OR (95%CI) = 2.40(1.35-4.23)), or T2D+ CAD (P = 1.14610 25 , OR (95%CI) = 3.38 (1.95-5.87)) compared with Control individuals. Finally, we calculated the genetic dosedependent cumulative risk, by comprising those cases and controls who harbored two or more than three risk alleles with carriers of one or no risk alleles. Significant P trend values were obtained for all Variants in ADIPOR1 Associated with T2D and CAD PLOS ONE | www.plosone.org three disease groups: T2D (P trend = 1.79610 24 ), CAD (P trend = 0.002), and T2D+CAD (P trend = 1.35610 25 ). (Table 4).

Discussion
Using a multi-case-control comparison in a northern Chinese population, we identified two variants in the ADIPOR1 gene, which supported our hypothesis that the etiology of CAD, T2D, and T2D with CAD was partially associated with ADIPOR1 polymorphisms. rs3737884, G was shown to be associated with an increased risk of all three diseases, while rs16850797, C was associated with T2D and T2D with CAD. To the best of our knowledge, neither polymorphism has previously been reported in association with diseases.
Interestingly, we also found that the risk genotypes of both rs3737884 and rs16850797 were consistently associated with five common metabolic phenotypes in T2D, CAD, and T2D with CAD. Thus, it is conceivable that the ADIPOR1 risk variants are not only shared by all three diseases status, but also by different metabolic phenotypes. This would result in a simplified, more consistent association between ADIPOR1 and the three diseases status. Therefore, we postulate that the ADIPOR1 variants act on the development of other metabolic disturbances such as IR, which partially contributes to the etiology of T2D, CAD, and T2D with CAD by fulfilling the ''common soil'' hypothesis.
Previously, Soccio et al. showed that three SNPs (including rs7539542) out of six tag SNPs selected from the HapMap-CEU database were significantly associated with CAD susceptibility among individuals with T2D [13]. Based on their analysis, when using major allele homozygote as the reference between CADpositive and-negative T2D subjects in our data, we found that the GC genotype of rs7539542 was possibly, but not convincingly associated with an increased risk of CAD (GC vs. CC, P = 0.043, OR (95%CI) = 1.20(1.02-3.87); GG vs. CC P = 0.617, OR (95%CI) = 1.19(0.60-2.37)). Although the risk allele in both cases was G, it was the major allele in our analysis and the minor allele in the study by Soccio et al. [13]. Based on HapMap and Ensembl databases, we found that the allelic distribution of rs7539542 in the subjects of our study was similar to that of HCB and CHS, but significantly different from CEU (Table S4). This suggests that the different genetic background of various ethnic populations might affect CAD susceptibility and could be more informative for rs7539542 in European than in Asian populations.
We also found that the frequencies of rs3737884, G did not differ significantly between any of the disease groups, and that the frequencies of rs16850797, C were not significantly different between T2D and T2D+CAD groups. This case-only analysis therefore showed that shared ADIPOR1 variants were not associated with disease status, which could reflect the fact that they posed a similar risk to the development of T2D, CAD, and T2D with CAD. It also indicates that rs3737884, G and rs16850797, C overlap slightly in their contribution to the etiology of these diseases, which further supports the ''common soil'' hypothesis that diabetes and cardiovascular disease share common antecedents.
Phenotype analyses showed that the risk genotypes of rs3737884 (GG+AG) and rs16850797 (CC+CG) were consistently associated with common metabolic phenotypes which were considered to be risk factors for the three diseases (Table S6). These were higher BMI, SBP, DBP, FBG, and lower HDL-C levels, which showed significant differences in the disease groups compared with the Control group. Our findings could prove the previously identified relationship between adiponectin and metabolic traits [16], which was also observed in animal models [4,18]. ADIPOR1 knockout Table 2. Association of SNPs in ADIPOR1 with T2D, CAD and T2D with CAD in additive genetic model. mice had significantly impaired glucose tolerance and significantly higher plasma insulin concentrations compared with wild-type [4]. By contrast, macrophage-specific ADIPOR1 transgenic mice showed reductions in whole body weight, TC, TG, and free fatty acid, and improved glucose tolerance and insulin sensitivity [18]. Because these metabolic traits are considered traditional risk factors for CAD, and as individuals with T2D often show these metabolic abnormalities, this indicates that T2D and CAD may share partial genetic susceptibility, which could include ADIPOR1 polymorphisms. However, further functional pathogenesis studies are needed to confirm this. Although some metabolic phenotypes are common to T2D, CAD, and T2D with CAD, there are few reported genetic variants shared by the three diseases. The present study identified such variants in ADIPOR1 and our genotype-phenotype study showed that these variants were also associated with common metabolic phenotypes in T2D, CAD, and T2D with CAD. Thus, it is possible that the shared variants have a more extensive pathophysiological impact on disease, common metabolic phenotypes, and functional disturbance in the human body. Moreover, we speculate that the ''common soil'' of these diseases is in fact the common metabolic phenotypes and functional disturbances that lead to the development of T2D, CAD, and T2D with CAD. The risk genotypes shared by all three diseases could therefore contribute to disease susceptibility.

SNPs
The observed trend of increased dose-dependent cumulative genetic risk with increasing risk allele numbers indicates that the risk alleles might be a form of quantity trait loci (QTL) that contributes to the risk of disease occurrence. Because genetic predisposition to complex disease is thought to reflect the cumulative effect of variants, our results suggest that susceptibility to T2D, CAD, or T2D with CAD is dose-dependent with  Table 4. Cumulative effect analyses of rs3737884 and rs16850797 in ADIPOR1 associated with risk of T2D, CAD and T2D with CAD. increasing numbers of risk alleles. This could be explained by speculating that a higher number of risk variants have a potentially synergistic effect on interference with the interaction or affinity between receptors and ligands, affecting events within target cells such as signal transduction. Repeating our analysis on a larger cohort would be useful to confirm the associations observed here in a relatively small sample size and help us to better understand the underlying pathogenesis of the three diseases.
In summary, we identified ADIPOR1 SNPs rs3737884 and rs16850797 as shared genetic variants that were consistently associated with T2D, CAD, and T2D with CAD in a northern Chinese population. Our data suggests that ADIPOR1 polymorphisms have a QTL risk and cause pleiotropic effects that occur during the development of the three diseases. These findings could provide novel insights into the etiology of metabolic diseases as well as the development of genetic markers to identify these diseases in a clinical setting. Table S1 Basic information for genotyping. F: forward primer; R: reverse primer; P: probe; HRM: high-resolution melting; RFLP: restriction fragment length polymorphism. *rs12045862 was genotyped using unlabeled probe method of HRM. **R:A or G;Y:C or T;N:A,C,G or T. (DOC) Table S2 The multiple comparison results of clinical and demographic characteristics among groups. Variables were expressed as percentage, mean 6 standard deviation, or median (interquartile range); Differences between characteristics were compared using parametric (Student's t-test for normally distributed variables) or nonparametric (Mann-Whitney U test for non-normally distributed variables) methods for continuous variables, and Pearson's x 2 analysis for categorical variables. CAD, coronary artery disease; T2D, type 2 diabetes; T2D+CAD: T2D with CAD; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG, fasting plasma glucose; TC, total cholesterol; TG, triglycerides; BMI, body mass index. Six independent hypotheses are tested with a significance level set at P#0.008 (0.05/6) according to Bonferroni correction.

(DOC)
Table S3 Allelic distribution of the 6 SNPs in ADIPOR1 in our study. SNPs: single nucleotide polymorphism; CAD, coronary artery disease; T2D, type 2 diabetes; T2D+CAD: T2D with CAD. *These alleles were defined on the basis of the alleles contrast in this study. **The former is ancestral allele. *** UTR-3: untranslated region. (DOC) Table S5 Association of rs3737884 and rs16850797 with diseases in three types of genetic models. CAD, coronary artery disease; T2D, type 2 diabetes; T2D+CAD: T2D with CAD; OR, odds ratios; CI, confidence interval. All OR and P values are obtained by Pearson's x 2 or unconditional logistic regression and adjusted for gender, age and body mass index. All variants with nominal significance (P#0.05) are listed; the threshold for significance by Bonferroni correction is 0.05/3 = 0.017 (three independent hypotheses: T2D vs. Control, CAD vs. Control, T2D with CAD vs. Control).*P value that can pass multiple testing correction (P#0.017); E indicates the power of the base-10 exponent (i.e. 9.80E-05 = 9.80610 25 ). The df of a per-allele OR value is 2 in the additive genetic model analysis. ORs are computed using wild homozygous carriers of variant as the reference group in the dominance model analysis and non-risk homozygous carriers of variant as the reference group in the recessive model. The risk alleles are rs3737884, G and rs16850797, C, respectively. (DOC) Table S6 Covariates analyses of risk genotypes in ADIPOR1 associated with risk of T2D, CAD and T2D with CAD. Variables were expressed as percentage, mean 6 standard deviation, or median (interquartile range). Risk genotypes were defined on the basis of the genetic dominance model analysis in Table S5. Differences between covariates were compared using parametric (Student's t-test for normally distributed variables) or nonparametric (Mann-Whitney U test for nonnormally distributed variables) methods for continuous variables, and P values are obtained by Pearson's x 2 for categorical variables. *Statistical significances are considered as P#0.05; ** Statistical significances are considered as P#0.017 according to Bonferroni correction under three independent hypotheses. CAD, coronary artery disease; T2D, type 2 diabetes; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; FBG, fasting plasma glucose; TG, triglycerides; TC, total cholesterol; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol. (DOC)