OBJECTIVE

Adiponectin, a hormone secreted by adipose tissue, is of particular interest in metabolic syndrome, because it is inversely correlated with obesity and insulin sensitivity. However, it is not known to what extent the genetics of plasma adiponectin and the genetics of obesity and insulin sensitivity are interrelated. We aimed to evaluate the heritability of plasma adiponectin and its genetic correlation with the metabolic syndrome and metabolic syndrome–related traits and the association between these traits and 10 ADIPOQ single nucleotide polymorphisms (SNPs).

RESEARCH DESIGN AND METHODS

We made use of a family-based population, the Erasmus Rucphen Family study (1,258 women and 967 men). Heritability analysis was performed using a polygenic model. Genetic correlations were estimated using bivariate heritability analyses. Genetic association analysis was performed using a mixed model.

RESULTS

Plasma adiponectin showed a heritability of 55.1%. Genetic correlations between plasma adiponectin HDL cholesterol and plasma insulin ranged from 15 to 24% but were not significant for fasting glucose, triglycerides, blood pressure, homeostasis model assessment of insulin resistance (HOMA-IR), and C-reactive protein. A significant association with plasma adiponectin was found for ADIPOQ variants rs17300539 and rs182052. A nominally significant association was found with plasma insulin and HOMA-IR and ADIPOQ variant rs17300539 after adjustment for plasma adiponectin.

CONCLUSIONS

The significant genetic correlation between plasma adiponectin and HDL cholesterol and plasma insulin should be taken into account in the interpretation of genome-wide association studies. Association of ADIPOQ SNPs with plasma adiponectin was replicated, and we showed association between one ADIPOQ SNP and plasma insulin and HOMA-IR.

The dramatic increase in the prevalence of the metabolic syndrome in countries with a western lifestyle is precipitated by environmental variables. However, the individual susceptibility to the obesogenic environment is largely determined by genetic susceptibility (1). Central obesity, dyslipidemia, impaired glucose metabolism, and hypertension are the key elements determining the expression of the metabolic syndrome (2), which is associated with an increased risk for type 2 diabetes and cardiovascular disease (2).

Adipose tissue is an active endocrine tissue that can respond to changes in metabolic conditions by secreting biologically active substances (adipokines). The adipokine family can be divided into two overlapping sets of signaling molecules, namely those with metabolic/immunological function, which include interleukins 1β, 6, 8, 10, or 18, tumor necrosis factor-α and transforming growth factor-β, and those with endocrine function, which include leptin, retinol-binding protein 4, adiponectin, and resistin (3). Human adiponectin is a protein of 247 amino acids (30-kDa), encoded by a gene (ADIPOQ) located on chromosome 3q27 (4). Adiponectin is secreted and present in plasma in various multimeric forms, for which the biological significance remains to be determined. Rasmussen-Torvik et al. (5) showed that binding of adiponectin to adiponectin receptors (ADIPOR1 and ADIPOR2) in mice results in increased AMP-activated protein kinase activity and peroxisome proliferator–activated receptor-α activity. In humans, both receptors are expressed mainly in skeletal muscle and adiponectin could thus play a role in energy metabolism.

The role of adiponectin in energy metabolism is confirmed by its inverse correlation with body weight, metabolic syndrome, metabolic syndrome–related traits, and type 2 diabetes (6). In mouse models, adiponectin has been shown to play a role in energy homeostasis by regulating insulin sensitivity of the liver (7). In addition, adiponectin is suggested to exhibit anti-inflammatory properties (8). Thus, adiponectin could play a role in obesity–induced impairment of the metabolic state, systemic inflammation, and the corresponding risk for cardiovascular disease.

Limited data on the overall heritability of plasma adiponectin are available. Furthermore, it is not known whether the genetics of plasma adiponectin overlap with the genetics of body weight and insulin sensitivity/diabetes or other individual components of the metabolic syndrome. Several studies showed convincing association of genetic variants near and in the promoter region of the ADIPOQ gene with plasma adiponectin and type 2 diabetes or type 2 diabetes–related traits (5,9,10).

In the present study, we set out to evaluate the heritability of plasma adiponectin and its genetic correlation with the metabolic syndrome and metabolic syndrome–related traits (BMI, insulin, homeostasis model assessment of insulin resistance [HOMA-IR], and plasma C-reactive protein [CRP]).

In the present study, we used data of the Erasmus Rucphen Family (ERF) study, which is embedded into a rural genetically isolated population (Genetic Research in Isolated Populations [GRIP]). This young, genetic isolate from the southwest Netherlands was initiated by <400 founders in the middle of 18th century. Minimal immigration occurred among the surrounding settlements for social and religious reasons. The population experienced a fast expansion and at the moment this region includes roughly 20,000 inhabitants. The ERF population is a cross-sectional cohort and includes 3,000 individuals who were not selected based on health information but rather comprise living descendants of 22 couples who had at least six children baptized in the community church in approximately 1850–1900. Details about the genealogy of the population have been published elsewhere (11,12). In the current study, we included 2,256 individuals of the ERF population for whom all study parameters were known. We did not exclude participants based on health status. The study protocol was approved by the medical ethics board of the Erasmus MC Rotterdam (Rotterdam, Netherlands). All investigations were carried out in accordance with the Declaration of Helsinki.

Data collection

Blood from participants was obtained in a fasted state. Total plasma insulin measurements were analyzed with an INS-IRMA kit (BioSource), total plasma adiponectin with a human adiponectin RIA kit (Linco Research), and total plasma CRP with a CRP ELISA (Diagnostic Systems Laboratories). All measurements were performed to conform to the manufacturer's protocol. Insulin sensitivity was based on the HOMA-IR) (glucose × insulin/22.5). Plasma CRP showed kurtosis; therefore, upper plasma CRP levels exceeding three times the SD of the mean were removed from further analyses. Data on plasma adiponectin, insulin, HOMA-IR, and CRP were available for 2,256 individuals. Metabolic syndrome was assessed according to the criteria of the International Diabetes Federation (2006, Europids), which requires a minimum waist circumference and two of the following abnormalities: high fasting plasma glucose, low HDL cholesterol, high total plasma triglycerides, and high systolic (SBP) and/or diastolic blood pressure (DBP), described in detail in supplementary Table 1 (available in an online appendix at http://care.diabetesjournals.org/cgi/content/full/dc09-1385/DC1). Prevalence and heritability of the metabolic syndrome and related traits in the ERF have been reported previously (1,13).

Statistical analyses

Analysis of mean differences between groups was tested using ANOVA statistics (continuous variables) or χ2 statistics (categorical variables). Correlations of plasma adiponectin with other traits were estimated in men and women separately with adjustment for age and BMI. Heritability estimations were obtained using SOLAR software (version 2.05; http://solar.sfbrgenetics.org) (1,11). Heritability and genetic association analyses were performed using normal log-transformed trait values and the “tdist” function in SOLAR. The polygenic model (covariates: sex and age) was applied. Heritability estimations included a second variance component, the sibship effect, which is an estimate of phenotypic similarity, due to effects of a shared (early) environment and genetically dominant effects (1). The effect of relevant medication use was assumed to be covered by the metabolic syndrome definition (model B). Inbreeding coefficients in heritability estimations were not significant and therefore were excluded from further analyses. To determine the genetic correlations of plasma adiponectin with other traits we used bivariate heritability analysis. We also estimated sex-specific heritability using bivariate analysis of traits stratified by sex. Because the latter analysis is confined to one quantitative trait, the environmental correlation component is forced to be zero. Bivariate analysis was applied to plasma adiponectin with the metabolic syndrome components and metabolic syndrome–related components using sex, age, and BMI as covariates (model A) or using sex, age, and metabolic syndrome as covariates (model B). The bivariate heritability analysis yields estimates of the total overlapping genetic and environmental component (correlation) of these traits.

The following ADIPOQ variants were selected after literature review (9) and confirmed for tag property using Haploview (r2 > 0.8; minor allele frequency >10%) and CEU HapMap data: rs864265, rs822387, rs17300539 (−11391G/A), rs266729, rs182052, rs822396, rs2241766 (+45T/G), rs1501299 (+276G/T), rs3774262, and rs6773957. Selected variants were genotyped using a Sequenom iPLEX application (matrix-assisted laser desorption ionization/time of flight; Sequenom, San Diego, CA). Genotypes were screened for Mendelian errors using the pedigree structure (14). All 10 ADIPOQ variants achieved a call rate of >95% and all were in Hardy-Weinberg equilibrium (P > 0.05). For ADIPOQ genotype association, we assumed an additive model. Analysis of plasma adiponectin included the covariates sex, age, and BMI (model A) or covariates sex, age, and metabolic syndrome (model B). To investigate whether ADIPOQ single nucleotide polymorphisms (SNPs) were independently associated with plasma adiponectin, we applied a backward linear regression model containing all SNPs that were associated with plasma adiponectin at a false discovery rate <0.05. The following model was used: rs822387, rs17300539, rs182052, rs1501299, rs6773957, sex, age, and BMI as independent variables and plasma adiponectin as the dependent variable. The independence of the associated SNPs was confirmed in Haploview, and these SNPs were used in association analysis with the metabolic syndrome–related traits.

Association analyses of ADIPOQ variants with metabolic syndrome–related traits was performed using two models with different covariates: 1) sex, age, and BMI and 2) sex, age, BMI, and plasma adiponectin. Adjustment for family structure of the association model was based on a pedigree matrix obtained using Illumina 6K linkage chip data. Analysis was performed using model residuals in a score test accounting for pedigree structure as implemented in GenABEL software (15) function “mmscore” (16). The Bayesian information criterion (BIC) for plasma adiponectin was implemented on associated ADIPOQ SNPs, sex, age, and BMI using R software. All other analyses were performed using SPSS 14.01 (September 2005; SSPS, Chicago, IL) software.

Characteristics of the ERF cohort are presented in Table 1. The prevalence of the metabolic syndrome, according to the International Diabetes Federation definition, was 29.3% in women and 37.1% in men. Virtually all mean values differed significantly (P < 0.01) between sexes.

Table 1

Characteristics of the ERF population

WomenMen
n 1,258 967 
Age (years) 48.0 ± 14.3 49.4 ± 14.1* 
    Range (years)/minimum–maximum 68.0/18–86 68.5/18–86 
Metabolic syndrome and its individual components   
    Metabolic syndrome 369 (29.3) 359 (37.1) 
    Waist circumference (cm) 81.9 ± 12.0 94.2 ± 11.5 
    Glucose (mmol/l) 4.4 ± 0.9 4.8 ± 1.0 
    HDL cholesterol (mmol/l) 1.4 ± 0.4 1.1 ± 0.3 
    Triglycerides (mmol/l) 1.2 ± 0.7 1.5 ± 0.9 
    SBP (mmHg) 136.1 ± 21.2 143.4 ± 18.0 
    DBP (mmHg) 78.6 ± 9.8 81.7 ± 9.8 
Metabolic syndrome–related traits   
    BMI (kg/m226.6 ± 5.0 27.3 ± 4.1 
    Insulin (μU/ml) 12.8 ± 5.8 13.9 ± 8.8* 
    HOMA-IR 2.6 ± 1.6 3.0 ± 2.4 
    Adiponectin (mg/l) 12.4 ± 5.8 8.1 ± 4.1 
    CRP (mg/l) 11.1 ± 22.2 10.0 ± 28.5 
Medication   
    Glucose-lowering 26 (2.1) 27 (2.8) 
    Lipid-lowering 132 (10.5) 144 (14.9)* 
    Antihypertension 237 (18.8) 201 (20.8) 
WomenMen
n 1,258 967 
Age (years) 48.0 ± 14.3 49.4 ± 14.1* 
    Range (years)/minimum–maximum 68.0/18–86 68.5/18–86 
Metabolic syndrome and its individual components   
    Metabolic syndrome 369 (29.3) 359 (37.1) 
    Waist circumference (cm) 81.9 ± 12.0 94.2 ± 11.5 
    Glucose (mmol/l) 4.4 ± 0.9 4.8 ± 1.0 
    HDL cholesterol (mmol/l) 1.4 ± 0.4 1.1 ± 0.3 
    Triglycerides (mmol/l) 1.2 ± 0.7 1.5 ± 0.9 
    SBP (mmHg) 136.1 ± 21.2 143.4 ± 18.0 
    DBP (mmHg) 78.6 ± 9.8 81.7 ± 9.8 
Metabolic syndrome–related traits   
    BMI (kg/m226.6 ± 5.0 27.3 ± 4.1 
    Insulin (μU/ml) 12.8 ± 5.8 13.9 ± 8.8* 
    HOMA-IR 2.6 ± 1.6 3.0 ± 2.4 
    Adiponectin (mg/l) 12.4 ± 5.8 8.1 ± 4.1 
    CRP (mg/l) 11.1 ± 22.2 10.0 ± 28.5 
Medication   
    Glucose-lowering 26 (2.1) 27 (2.8) 
    Lipid-lowering 132 (10.5) 144 (14.9)* 
    Antihypertension 237 (18.8) 201 (20.8) 

Data are means ± SD for continuous traits and n (%) for categorical traits.

*Significantly different from women (P < 0.05).

P < 0.01, continuous traits based on ANOVA and noncategorical traits based on χ2 test.

To determine the association of plasma adiponectin with the metabolic syndrome and related traits, we calculated partial correlation coefficients of plasma adiponectin with each trait studied (supplementary Table 2, available in an online appendix) adjusted for age and BMI. Plasma adiponectin showed a high correlation with the metabolic syndrome (ρ = −0.20 and ρ = −0.13 in women and men, respectively) and with the lipid components HDL cholesterol and triglycerides (all ρ ≥ 0.15, P < 0.01). The correlation of plasma adiponectin with BMI (adjustment only for age), fasting plasma insulin, HOMA-IR, and plasma CRP showed correlations coefficients ranging from 0.10 to 0.29 (all P < 0.05). In men, we observed, in general, lower correlation coefficients for all components, with the exception of DBP.

We next estimated the heritability of plasma adiponectin and metabolic syndrome–related traits. Heritability estimates of the metabolic syndrome and its individual components and BMI in the ERF were described earlier (1,13). All heritability estimates were highly significant and none of the metabolic syndrome–related traits showed a significant sibship effect estimate (S). The heritability estimate of plasma insulin was 21.4%, SEM = 5.2%, P < 10−05 (S = 6.1%, SEM = 4.2%, P = 0.07), of HOMA-IR was 22.0%, SEM = 5.3%, P < 10−05 (S = 5.8%, SEM = 4.3%, P = 0.08), and of plasma CRP level was 21.0%, SEM = 4.9%, P < 10−05 (S = nil). The highest heritability estimate was found for plasma adiponectin: 55.1%, SEM = 4.7, P < 10−6 (S = nil).

Next, we performed bivariate analysis of sex-stratified age-adjusted traits. We observed no significant difference in the heritability between the sexes and found a high correlation (ρ = 100%) between the heritability of plasma adiponectin in men (h2 = 59.6%, SEM = 9.4) and women (h2 = 52.9%, SEM = 7.7). Details on sex-specific heritability of plasma insulin, HOMA-IR, and plasma CRP are presented detail in supplementary Table 3 (available in an online appendix). Sex-stratified bivariate heritability estimates of the metabolic syndrome and its individual components and BMI were described earlier (1,13).

Bivariate heritability analyses were performed using trait-by-trait analyses on plasma adiponectin combined with the metabolic syndrome and metabolic syndrome–related traits. To investigate whether the interrelation between the traits affected their genetic correlations, we used two statistical models: model A included BMI as the covariate and model B included metabolic syndrome as the covariate in addition to age and sex in the trait-by-trait analyses. The outcomes of the bivariate heritability analysis of plasma adiponectin and the metabolic syndrome and the metabolic syndrome related traits are presented in Table 2.

Table 2

Genetic and environmental correlation of plasma adiponectin with metabolic syndrome, metabolic syndrome components, and metabolic syndrome–related traits

ρGSEMPρESEMP*
Model A       
    Metabolic syndrome and its individual components 
        Metabolic syndrome −42.7 11.3 <10−04 −19.4 9.4 0.04 
        Waist circumference −20.4 7.3 <10−03 −18.8 5.8 <10−04 
        Glucose −2.6 8.7 0.76 −12.0 5.6 0.03 
        HDL cholesterol 24.7 7.2 <10−04 35.6 5.5 <10−9 
        Triglycerides −12.2 8.9 0.17 −26.0 5.2 <10−6 
        SBP −10.3 9.1 0.26 −7.1 5.3 0.18 
        DBP −9.5 9.2 0.31 6.0 5.4 0.27 
    Metabolic syndrome–related traits       
        BMI −32.4 7.2 <10−05 −20.7 5.9 <10−04 
        Insulin −20.0 9.9 0.04 −18.8 5.1 <10−04 
        HOMA-IR −18.6 9.9 0.06 −19.6 5.1 <10−04 
        CRP −11.6 12.3 0.34 −13.4 5.4 0.01 
Model B       
    Metabolic syndrome individual components 
        Waist circumference −15.0 8.9 0.092 −22.6 5.6 <10−04 
        Glucose 0.5 8.9 0.955 −11.3 5.5 0.040 
        HDL cholesterol 15.2 7.7 0.048 37.2 5.4 <10−11 
        Triglycerides −11.9 9.8 0.225 −26.0 5.1 <10−06 
        SBP −6.3 9.3 0.498 8.5 5.3 0.100 
        DBP −7.1 9.6 0.459 4.0 5.3 0.450 
    Metabolic syndrome–related traits       
        BMI −12.9 7.9 0.103 −14.6 6.1 0.017 
        Insulin −19.1 10 0.056 −19.1 5.1 <10−03 
        HOMA-IR −16.4 10 0.101 −20.4 5.1 <10−04 
        CRP −12.3 11.7 0.293 −14.4 5.5 0.009 
ρGSEMPρESEMP*
Model A       
    Metabolic syndrome and its individual components 
        Metabolic syndrome −42.7 11.3 <10−04 −19.4 9.4 0.04 
        Waist circumference −20.4 7.3 <10−03 −18.8 5.8 <10−04 
        Glucose −2.6 8.7 0.76 −12.0 5.6 0.03 
        HDL cholesterol 24.7 7.2 <10−04 35.6 5.5 <10−9 
        Triglycerides −12.2 8.9 0.17 −26.0 5.2 <10−6 
        SBP −10.3 9.1 0.26 −7.1 5.3 0.18 
        DBP −9.5 9.2 0.31 6.0 5.4 0.27 
    Metabolic syndrome–related traits       
        BMI −32.4 7.2 <10−05 −20.7 5.9 <10−04 
        Insulin −20.0 9.9 0.04 −18.8 5.1 <10−04 
        HOMA-IR −18.6 9.9 0.06 −19.6 5.1 <10−04 
        CRP −11.6 12.3 0.34 −13.4 5.4 0.01 
Model B       
    Metabolic syndrome individual components 
        Waist circumference −15.0 8.9 0.092 −22.6 5.6 <10−04 
        Glucose 0.5 8.9 0.955 −11.3 5.5 0.040 
        HDL cholesterol 15.2 7.7 0.048 37.2 5.4 <10−11 
        Triglycerides −11.9 9.8 0.225 −26.0 5.1 <10−06 
        SBP −6.3 9.3 0.498 8.5 5.3 0.100 
        DBP −7.1 9.6 0.459 4.0 5.3 0.450 
    Metabolic syndrome–related traits       
        BMI −12.9 7.9 0.103 −14.6 6.1 0.017 
        Insulin −19.1 10 0.056 −19.1 5.1 <10−03 
        HOMA-IR −16.4 10 0.101 −20.4 5.1 <10−04 
        CRP −12.3 11.7 0.293 −14.4 5.5 0.009 

Data are % unless indicated otherwise. Model A included covariates sex, age, and BMI; model B included covariates sex, age, and metabolic syndrome.

*P values were derived using a χ2 test and covariates sex, age, and BMI.

†Covariates sex and age only.

According to model A, metabolic syndrome, waist circumference, and HDL cholesterol demonstrated a significant (P < 10−3) shared genetic component with plasma adiponectin (−42.7, −32.4, and +24.7%, respectively). Triglycerides, fasting plasma glucose, and both SBP and DBP showed a low and insignificant shared genetic component with plasma adiponectin. BMI and plasma insulin showed a significant (P < 0.04) shared genetic component with plasma adiponectin of, respectively, −32.4 and −20.0%, whereas our finding for HOMA-IR was borderline significant (genetic correlation [ρG] = −18.6%, P = 0.06). The genetic correlation of plasma adiponectin and plasma CRP was found to be not significant. The environmental correlations of plasma adiponectin with plasma glucose (−12.0%), HDL cholesterol (+35.6%), triglycerides (−26.0%), and plasma CRP (−13.4%) exceeded their shared genetic component value. In contrast, the genetic correlation of plasma adiponectin and metabolic syndrome was twice as high as their environmental correlation (ρG = −42.7% and environmental correlation [ρE] = −19.4%). According to model B (covariates sex, age, and metabolic syndrome), HDL cholesterol shared a significant (15.2%, P < 0.05) genetic component with plasma adiponectin. Plasma insulin shared a borderline significant (19.1%, P = 0.056) genetic component with plasma adiponectin. All observed environmental correlations obtained according to model B exceeded their genetic correlations (Table 2).

To investigate to what extent the high heritability of plasma adiponectin is due to variants located in the ADIPOQ gene, we performed genetic association analysis. Table 3 presents association of 10 ADIPOQ SNPs with plasma adiponectin, adjusted for sex, age, and BMI (model A) and adjusted for sex, age, and metabolic syndrome (model B). According to both models A and B, nominally significant associations with plasma adiponectin were found for two promoter SNPs (rs822387 and rs17300539), one SNP located in exon 1 (rs182052), one SNP locate in exon 2 (rs1501299), and one SNP (rs6773957) located in the 3′-untranslated region of the ADIPOQ gene. Four of these SNPs (except rs822396 and rs1501299) remained significant after conservative Bonferroni correction (P < 0.005). The effect size of the promoter SNPs was substantially higher than that of the other significantly associated SNPs.

Table 3

Association of variants in and around ADIPOQ with plasma adiponectin

SNPPositionnAlleles (m/M)Frequency (m)Effect (m)SEMP
Model A        
    rs864265 Pr 1,918 G/T 0.19 −0.012 −0.023 0.601 
    rs822387 Pr 1,917 C/T 0.95 0.146 0.040 3.0*10−4 
    rs17300539 Pr 1,919 A/G 0.95 0.155 0.040 9.3*10−5 
    rs266729 Pr 1,934 G/C 0.74 −0.025 −0.021 0.230 
    rs182052 Ex-1 1,916 A/G 0.68 −0.070 −0.020 3.0*10−4 
    rs822396 Ex-1 1,925 A/G 0.22 0.043 0.021 0.049 
    rs2241766 In-1 1,939 G/T 0.93 0.037 0.036 0.303 
    rs1501299 Ex-2 1,921 T/G 0.73 0.057 0.021 0.006 
    rs3774262 Ex-2 1,937 A/G 0.93 0.035 0.036 0.330 
    rs6773957 3′-UTR 1,914 A/G 0.67 0.061 0.020 0.002 
Model B        
    rs864265 Pr 1,904 G/T 0.19 −0.006 −0.023 0.788 
    rs822387 Pr 1,904 C/T 0.95 0.143 0.040 3.6*10−4 
    rs17300539 Pr 1,905 A/G 0.95 0.166 0.040 2.7*10−5 
    rs266729 Pr 1,920 G/C 0.74 −0.017 −0.020 0.400 
    rs182052 Ex-1 1,902 A/G 0.68 −0.066 −0.019 6.7*10−4 
    rs822396 Ex-1 1,911 A/G 0.22 0.043 0.022 0.047 
    rs2241766 In-1 1,924 G/T 0.93 0.046 0.036 0.202 
    rs1501299 Ex-2 1,907 T/G 0.73 0.055 0.021 0.007 
    rs3774262 Ex-2 1,923 A/G 0.93 0.046 0.036 0.200 
    rs6773957 3′-UTR 1,900 A/G 0.67 0.063 0.020 0.001 
SNPPositionnAlleles (m/M)Frequency (m)Effect (m)SEMP
Model A        
    rs864265 Pr 1,918 G/T 0.19 −0.012 −0.023 0.601 
    rs822387 Pr 1,917 C/T 0.95 0.146 0.040 3.0*10−4 
    rs17300539 Pr 1,919 A/G 0.95 0.155 0.040 9.3*10−5 
    rs266729 Pr 1,934 G/C 0.74 −0.025 −0.021 0.230 
    rs182052 Ex-1 1,916 A/G 0.68 −0.070 −0.020 3.0*10−4 
    rs822396 Ex-1 1,925 A/G 0.22 0.043 0.021 0.049 
    rs2241766 In-1 1,939 G/T 0.93 0.037 0.036 0.303 
    rs1501299 Ex-2 1,921 T/G 0.73 0.057 0.021 0.006 
    rs3774262 Ex-2 1,937 A/G 0.93 0.035 0.036 0.330 
    rs6773957 3′-UTR 1,914 A/G 0.67 0.061 0.020 0.002 
Model B        
    rs864265 Pr 1,904 G/T 0.19 −0.006 −0.023 0.788 
    rs822387 Pr 1,904 C/T 0.95 0.143 0.040 3.6*10−4 
    rs17300539 Pr 1,905 A/G 0.95 0.166 0.040 2.7*10−5 
    rs266729 Pr 1,920 G/C 0.74 −0.017 −0.020 0.400 
    rs182052 Ex-1 1,902 A/G 0.68 −0.066 −0.019 6.7*10−4 
    rs822396 Ex-1 1,911 A/G 0.22 0.043 0.022 0.047 
    rs2241766 In-1 1,924 G/T 0.93 0.046 0.036 0.202 
    rs1501299 Ex-2 1,907 T/G 0.73 0.055 0.021 0.007 
    rs3774262 Ex-2 1,923 A/G 0.93 0.046 0.036 0.200 
    rs6773957 3′-UTR 1,900 A/G 0.67 0.063 0.020 0.001 

Model A included covariates sex, age, and BMI; model B included covariates sex, age, and metabolic syndrome.

*P values based on χ2 test. P values in bold were selected for analysis in a backward linear regression model and were evaluated on linkage disequilibrium using Haploview (CEU HapMap). Ex, exon; IN, intron; m, effect allele; Pr, promoter; UTR, untranslated region.

To determine the best model explaining plasma adiponectin we used the BIC (17). The parameters used in the BIC analysis for plasma adiponectin were sex, age, BMI, and the ADIPOQ SNPs rs822387, rs17300539, rs182052, rs1501299, and rs6773957 (n = 1,914). BIC values are presented in supplementary Table 4 (available in an online appendix). The lowest BIC value was found for the model including four ADIPOQ SNPs (rs822387, rs17300539, rs182052, and rs6773957).

Backward linear regression analysis of the five associated SNPs indicated that rs17300539 and rs182052 were independently associated with plasma adiponectin (P < 0.001 and P < 0.004, respectively, in the joint model). Evaluation of r2 between SNPs using Haploview confirmed the independence of the two ADIPOQ SNPs (r2 = 0.2, plot not shown). Next, we analyzed whether the two significant SNPs, ADIPOQ rs17300539 and rs182052, were also associated with the metabolic syndrome and metabolic syndrome–related traits using two models. The first model included sex, age, and BMI as covariates. The second model included sex, age, BMI, and plasma adiponectin as covariates. The second model was used to investigate whether metabolic syndrome–related traits were associated independently of plasma adiponectin. No significant associations were found for rs182052. ADIPOQ rs17300539 showed a significant association with plasma insulin (n = 1919, β = 0.072, SEM = 0.031, P = 0.022) and with HOMA-IR (n = 1,892, β = 0.084, SEM = 0.036, P = 0.021), according to model 2.

Here we report a high heritability of plasma adiponectin (55.1%) and a similar genetic architecture between men and women. We also demonstrated that the genetic component of the HDL cholesterol and plasma insulin overlap significantly with that of plasma adiponectin. ADIPOQ rs17300539 and rs182052 were both found to contribute independently to the heritability of plasma adiponectin. Of these two SNPs, rs17300539 was also associated with plasma insulin and HOMA-IR, illustrating their genetic overlap with adiponectin.

Our estimate of the heritability of plasma adiponectin is similar to heritability (62%) reported by Patel et al. (18). Furthermore, our heritability estimates for the metabolic syndrome-related traits (plasma insulin, HOMA-IR, and plasma CRP) are in agreement with earlier reports (19,21). In our previous study (1) on the heritability of metabolic syndrome and its individual components, heritability varied from 10.6% (metabolic syndrome) to 42.9% (HDL cholesterol). The heritability of plasma adiponectin is higher than the heritability of metabolic syndrome or of its individual components, making it an attractive trait for genome-wide association (GWA) studies.

We did not find evidence for a sex-specific genetic component for plasma adiponectin (1,13). Patel et al. (18) studied the genetic correlation between plasma adiponectin and obesity traits in a large longitudinal family-based cohort and found that the genetic correlation (ρG) of plasma adiponectin with BMI and waist circumference was for both ∼−40%. This result matches well with our estimate of the genetic correlation between adiponectin and BMI of −32.4% using sex, age, and BMI as covariates (model A). Furthermore, we found, using model A, a high and significant shared genetic component between plasma adiponectin and the metabolic syndrome (−42.7%). Moreover, this genetic correlation was twice as large as the environmental correlation. Furthermore, according to model A, waist circumference (−20.4%), HDL cholesterol (+24.7%), and plasma insulin (−20.0%) shared a significant genetic component with plasma adiponectin. Because many of the study parameters are strongly associated with each other, we also studied the genetic correlations, adjusting for metabolic syndrome (model B). Applying this conservative model, the genetic correlations found according to model A were consistent, again revealing the genetic correlations between plasma adiponectin and HDL cholesterol and plasma insulin. Our findings imply that genetic studies of plasma adiponectin might also lead to the identification of genes associated with HDL cholesterol and plasma insulin.

The ADIPOQ gene has been found to be consistently and significantly associated with plasma adiponectin in genetic association studies (9,22). To investigate the role of the ADIPOQ gene in the genetic overlap between plasma adiponectin and metabolic syndrome traits, we performed association analysis using 10 ADIPOQ SNPs. Hivert et al. (9) showed a significant association of the ADIPOQ SNP rs17300539 with plasma adiponectin. Our study of 10 ADOPOQ SNPs showed that plasma adiponectin was significantly and consistently associated with 6 ADIPOQ variants using adjustment for sex, age, and BMI (model A) or adjustment for sex, age, and metabolic syndrome (model B). Moreover, our study of 10 ADOPOQ SNPs showed that plasma adiponectin was significantly and independently associated with both ADIPOQ rs17300539 and rs182052 variants. ADIPOQ rs17300539 is located in the promoter region of the ADIPOQ gene, whereas ADIPOQ rs182052 is located in exon 1 of the ADIPOQ gene. Whether these two variants are actually causal remains to be determined. Because our heritability and association analyses on plasma adiponectin and associated traits are in concert with findings of other studies in general cohorts or other genetically isolated populations, it seems unlikely that our findings are specific for the genetically isolated ERF cohort.

Plasma adiponectin is strongly associated with metabolic syndrome, obesity, and, in particular, with plasma insulin and insulin sensitivity. There is evidence indicating that insulin directly affects plasma adiponectin (23,25). Thus, it is likely that the genetics of plasma adiponectin and insulin overlap. Hivert et al. (9) reported that rs173766743 was associated with the incidence of type 2 diabetes. Our analyses showed that rs17300539 was also associated with plasma insulin and HOMA, independently of plasma adiponectin. We did not observe any effect on this association using metabolic syndrome instead of BMI as a covariate (data not shown). Because HOMA-IR is a measure for insulin resistance that reflects a pre-diabetic state, we analyzed the r2 between ADIPOQ rs17300539 and rs173766743 in Haploview using CEU HapMap data. This analysis did not show any evidence of linkage disequilibrium between these variants (r2 = 0). An explanation for this apparent discrepancy may be the adjustment for plasma adiponectin levels in our association analyses. Whether the association of rs173766743 with type 2 diabetes is independent of plasma adiponectin was not reported. The association of rs17300539 with plasma insulin and HOMA-IR is independent of plasma adiponectin in our analyses, which implies a direct effect of this SNP on plasma insulin and insulin sensitivity. One possible explanation is that ADIPOQ rs17300539 is associated with a functional variation of the adiponectin protein affecting insulin sensitivity independent of plasma levels of the protein. However, this association would have to be replicated in independent analyses before further investigation.

In summary, the present study confirms and extends the correlation of plasma adiponectin with HDL cholesterol and plasma insulin. The high heritability of plasma adiponectin is promising for GWAs. The genetics of plasma adiponectin is similar between sexes. The genetics of plasma adiponectin showed a significant and consistent overlap with the genetics of HDL cholesterol and plasma insulin, which implies that GWAs of plasma adiponectin might also result in the detection of genetic variation associated with HDL cholesterol and plasma insulin.

Genetic association analyses indicated that ADIPOQ variation is strongly associated with plasma adiponectin and also that ADIPOQ rs17300539 is associated with plasma insulin and HOMA-IR independently of plasma adiponectin. These genetic association data are thus in line with the observed genetic overlap between plasma adiponectin and plasma insulin.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

This study was financially supported by the Nutrigenomics Consortium and the Centre for Medical Systems Biology in the framework of the Netherlands Genomics Initiative.

No potential conflicts of interest relevant to this article were reported.

We thank the participants from Genetic Research in Isolated Populations, the Erasmus Rucphen Family, and the general practitioners and other clinicians who made this work possible. We thank Najaf Amin and Maksim Struchalin for their help with the genetic association analyses and Eline Slagboom for facilitating the ADIPOQ genotyping.

1.
Henneman
P
,
Aulchenko
YS
,
Frants
RR
,
van Dijk
KW
,
Oostra
BA
,
van Duijn
CM
.
Prevalence and heritability of the metabolic syndrome and its individual components in a Dutch isolate: the Erasmus Rucphen Family study
.
J Med Genet
2008
;
45
:
572
577
2.
Procopiou
M
,
Philippe
J
.
The metabolic syndrome and type 2 diabetes: epidemiological figures and country specificities
.
Cerebrovasc Dis
2005
;
20
(
Suppl. 1
):
2
8
3.
Fischer-Posovszky
P
,
Wabitsch
M
,
Hochberg
Z
.
Endocrinology of adipose tissue—an update
.
Horm Metab Res
2007
;
39
:
314
321
4.
Giannessi
D
,
Maltinti
M
,
Del Ry
S
.
Adiponectin circulating levels: a new emerging biomarker of cardiovascular risk
.
Pharmacol Res
2007
;
56
:
459
467
5.
Rasmussen-Torvik
LJ
,
Pankow
JS
,
Jacobs
DR
 Jr
,
Steinberger
J
,
Moran
A
,
Sinaiko
AR
.
The association of SNPs in ADIPOQ, ADIPOR1, and ADIPOR2 with insulin sensitivity in a cohort of adolescents and their parents
.
Hum Genet
2009
;
125
:
21
28
6.
Kadowaki
T
,
Yamauchi
T
,
Kubota
N
,
Hara
K
,
Ueki
K
,
Tobe
K
.
Adiponectin and adiponectin receptors in insulin resistance, diabetes, and the metabolic syndrome
.
J Clin Invest
2006
;
116
:
1784
1792
7.
Yamauchi
T
,
Kamon
J
,
Waki
H
,
Terauchi
Y
,
Kubota
N
,
Hara
K
,
Mori
Y
,
Ide
T
,
Murakami
K
,
Tsuboyama-Kasaoka
N
,
Ezaki
O
,
Akanuma
Y
,
Gavrilova
O
,
Vinson
C
,
Reitman
ML
,
Kagechika
H
,
Shudo
K
,
Yoda
M
,
Nakano
Y
,
Tobe
K
,
Nagai
R
,
Kimura
S
,
Tomita
M
,
Froguel
P
,
Kadowaki
T
.
The fat-derived hormone adiponectin reverses insulin resistance associated with both lipoatrophy and obesity
.
Nat Med
2001
;
7
:
941
946
8.
Wolf
AM
,
Wolf
D
,
Rumpold
H
,
Enrich
B
,
Tilg
H
.
Adiponectin induces the anti-inflammatory cytokines IL-10 and IL-1RA in human leukocytes
.
Biochem Biophys Res Commun
2004
;
323
:
630
635
9.
Hivert
MF
,
Manning
AK
,
McAteer
JB
,
Florez
JC
,
Dupuis
J
,
Fox
CS
,
O'Donnell
CJ
,
Cupples
LA
,
Meigs
JB
.
Common variants in the adiponectin gene (ADIPOQ) associated with plasma adiponectin levels, type 2 diabetes, and diabetes-related quantitative traits: the Framingham Offspring Study
.
Diabetes
2008
;
57
:
3353
3359
10.
Heid
IM
,
Wagner
SA
,
Gohlke
H
,
Iglseder
B
,
Mueller
JC
,
Cip
P
,
Ladurner
G
,
Reiter
R
,
Stadlmayr
A
,
Mackevics
V
,
Illig
T
,
Kronenberg
F
,
Paulweber
B
.
Genetic architecture of the APM1 gene and its influence on adiponectin plasma levels and parameters of the metabolic syndrome in 1,727 healthy Caucasians
.
Diabetes
2006
;
55
:
375
384
11.
Pardo
LM
,
Mackay
I
,
Oostra
B
,
van Duijn
CM
,
Aulchenko
YS
.
The effect of genetic drift in a young genetically isolated population
.
Ann Hum Genet
2005
;
69
:
288
295
12.
Aulchenko
YS
,
Heutink
P
,
Mackay
I
,
Bertoli-Avella
AM
,
Pullen
J
,
Vaessen
N
,
Rademaker
TA
,
Sandkuijl
LA
,
Cardon
L
,
Oostra
B
,
van Duijn
CM
.
Linkage disequilibrium in young genetically isolated Dutch population
.
Eur J Hum Genet
2004
;
12
:
527
534
13.
Zillikens
MC
,
Yazdanpanah
M
,
Pardo
LM
,
Rivadeneira
F
,
Aulchenko
YS
,
Oostra
BA
,
Uitterlinden
AG
,
Pols
HA
,
van Duijn
CM
.
Sex-specific genetic effects influence variation in body composition
.
Diabetologia
2008
;
51
:
2233
2241
14.
O'Connell
JR
,
Weeks
DE
.
PedCheck: a program for identification of genotype incompatibilities in linkage analysis
.
Am J Hum Genet
1998
;
63
:
259
266
15.
Aulchenko
YS
,
Ripke
S
,
Isaacs
A
,
van Duijn
CM
.
GenABEL: an R library for genome-wide association analysis
.
Bioinformatics
2007
;
23
:
1294
1296
16.
Chen
WM
,
Abecasis
GR
.
Estimating the power of variance component linkage analysis in large pedigrees
.
Genet Epidemiol
2006
;
30
:
471
484
17.
Ball
RD
.
Quantifying evidence for candidate gene polymorphisms: Bayesian analysis combining sequence-specific and quantitative trait loci colocation information
.
Genetics
2007
;
177
:
2399
2416
18.
Patel
SR
,
Larkin
EK
,
Redline
S
.
Shared genetic basis for obstructive sleep apnea and adiposity measures
.
Int J Obes (Lond)
2008
;
32
:
795
800
19.
Rasmussen-Torvik
LJ
,
Pankow
JS
,
Jacobs
DR
,
Steffen
LM
,
Moran
AM
,
Steinberger
J
,
Sinaiko
AR
.
Heritability and genetic correlations of insulin sensitivity measured by the euglycaemic clamp
.
Diabet Med
2007
;
24
:
1286
1289
20.
Pilia
G
,
Chen
WM
,
Scuteri
A
,
Orru
M
,
Albai
G
,
Dei
M
,
Lai
S
,
Usala
G
,
Lai
M
,
Loi
P
,
Mameli
C
,
Vacca
L
,
Deiana
M
,
Olla
N
,
Masala
M
,
Cao
A
,
Najjar
SS
,
Terracciano
A
,
Nedorezov
T
,
Sharov
A
,
Zonderman
AB
,
Abecasis
GR
,
Costa
P
,
Lakatta
E
,
Schlessinger
D
.
Heritability of cardiovascular and personality traits in 6,148 Sardinians
.
PLoS Genet
2006
;
2
:
e132
21.
Hsueh
WC
,
Silver
KD
,
Pollin
TI
,
Bell
CJ
,
O'Connell
JR
,
Mitchell
BD
,
Shuldiner
AR
.
A genome-wide linkage scan of insulin level derived traits: the Amish Family Diabetes Study
.
Diabetes
2007
;
56
:
2643
2648
22.
Ling
H
,
Waterworth
DM
,
Stirnadel
HA
,
Pollin
TI
,
Barter
PJ
,
Kesaniemi
YA
,
Mahley
RW
,
McPherson
R
,
Waeber
G
,
Bersot
TP
,
Cohen
JC
,
Grundy
SM
,
Mooser
VE
,
Mitchell
BD
.
Genome-wide linkage and association analyses to identify genes influencing adiponectin levels: the GEMS study
.
Obesity (Silver Spring)
2009
;
17
:
737
744
23.
Möhlig
M
,
Wegewitz
U
,
Osterhoff
M
,
Isken
F
,
Ristow
M
,
Pfeiffer
AF
,
Spranger
J
.
Insulin decreases human adiponectin plasma levels
.
Horm Metab Res
2002
;
34
:
655
658
24.
Hung
J
,
McQuillan
BM
,
Thompson
PL
,
Beilby
JP
.
Circulating adiponectin levels associate with inflammatory markers, insulin resistance and metabolic syndrome independent of obesity
.
Int J Obes (Lond)
2008
; 
32
:
772
779
25.
Brame
LA
,
Considine
RV
,
Yamauchi
M
,
Baron
AD
,
Mather
KJ
.
Insulin and endothelin in the acute regulation of adiponectin in vivo in humans
.
Obes Res
2005
;
13
:
582
588
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/licenses/by-nc-nd/3.0/ for details.

Supplementary data