Variations in Adiponectin Receptor Genes and Susceptibility to Type 2 Diabetes in Women

RESEARCH DESIGN AND METHODS

The Nurses’ Health Study was established in 1976 when 121,700 female registered nurses aged 30–55 years and residing in 11 large U.S. states completed a mailed questionnaire on medical history and lifestyle (18). The lifestyle factors, including smoking, menopausal status and postmenopausal hormone therapy, and body weight, have been updated by validated questionnaires every 2 years. Samples for the present case-control study were selected from a subcohort of 32,826 women who provided a blood sample between 1989 and 1990 and were free from diabetes, cardiovascular disease, stroke, or cancer at the time of blood collection. Incident cases were defined as self-reported diabetes confirmed by a validated supplementary questionnaire and diagnosed at least 1 year after blood collection through 2000. The supplementary questionnaire obtained information on symptoms, diagnostic tests, and hypoglycemic therapy used to define type 2 diabetic patients. Medical record review confirmed the diagnosis of type 2 diabetes with this questionnaire for 98% of cases, using the National Diabetes Data Group criteria (19). We used the American Diabetes Association diagnostic criteria for diagnosis of diabetes during the 1998 and 2000 cycles (20). In the present study, 714 incident cases of type 2 diabetes and 1,120 healthy control subjects were included. The selection of case and control subjects was described in detail elsewhere (21).

Assessment of covariates.

Anthropometric data and lifestyle factors were derived from the 1990 questionnaire. BMI was calculated as weight in kilograms divided by the square of height in meters. Physical activity was expressed as MET (metabolic equivalent task) hours based on self-reported types and durations of activities over the previous year.

Tagging SNPs selection and genotype determination.

DNA was extracted from the buffy coat fraction of centrifuged blood using the QIAmp Blood Kit (Qiagen, Chatsworth, CA). Tagging SNPs for ADIPOR1 were provided by A.D., who made the selection based on typing 28 common SNPs (minor allele frequency [MAF] ≥5%) covering the entire gene plus 5 Kb on each side (22). Tagging SNPs for ADIPOR2 were selected from HapMap (HapMap Public Release no. 19) using multimarker (aggressive) tagging mode (23). We also included several polymorphisms that were previously associated with diabetes risk. Polymorphisms were genotyped using TaqMan SNP allelic discrimination by means of an ABI 7900HT (Applied Biosystems, Foster City, CA). One polymorphism (rs4950894) had a relatively high missing value in genotyping (9.9%) and was not included in the analyses. The genotyped polymorphisms are presented in Table 1. Replicate quality control samples (10%) were included and genotyped with >99% concordance.

Statistical analyses.

A χ² test was used to assess whether the genotypes were in Hardy-Weinberg equilibrium and to compare the genotype and allele frequencies between case and control subjects. Odds ratios (ORs) were calculated using unconditional logistic regression adjusting for type 2 diabetes risk factors, including age (continuous), BMI (<23, 23–24.9, 25–29.9, 30–34.9, or ≥35 kg/m²), physical activity (<1.5, 1.5–5.9, 6.0–11.9, 12–20.9, and ≥21.0 MET h/week), smoking (never, past, and current), alcohol intake (nondrinker and drinker [0.1–4.9, 5–10, or >10 g/day]), family history of diabetes, and menopausal status (pre- or postmenopausal [never, past, or current hormone use]). As we simultaneously examined multiple gene polymorphisms, we used permutation testing to address the issue of multiple comparisons and to guide interpretation of nominally statistically significant associations. We estimated haplotype-specific ORs using an expectation-substitution approach to account for haplotype uncertainty given unphased genotype data (24,25). To test the global associations, we used a likelihood ratio test comparing a model with additive effects on the log odds scale for each common haplotype (treating the most common haplotype as the referent) to the intercept-only model. We considered haplotypes with greater than 5% frequency in at least one cohort or ethnic group to be “common.” The SAS statistical package was used for the analyses (SAS, version 8.2 for UNIX). All P values are two sided.

RESULTS

The location and allele frequency of examined ADIPOR1 and ADIPOR2 polymorphisms is presented in Table 1. Among the study participants, all polymorphisms conformed to Hardy-Weinberg equilibrium (P > 0.05). The MAF of ADIPOR1 polymorphisms ranges from 0.28 to 0.46, whereas the MAF of ADIPOR2 polymorphisms ranges from 0.07 to 0.48. Figure 1 shows the LD matrix for ADIPOR1 and ADIPOR2 polymorphisms. The two 3′ UTR polymorphisms rs1139646 and rs10920531 are in strong LD (D′ = 0.93; r² = 0.68) while the other four polymorphisms appeared to be in strong LD (D′ ≥ 0.97; r² ranges from 0.15 to 0.99). There were broadly three LD blocks in the ADIPOR2 gene (Fig. 1). The polymorphisms in ADIPOR1 and ADIPOR2 genes were not associated with BMI.

Polymorphism rs1139646 at 3′ UTR of ADIPOR1 showed nominally significant association with an increased risk of type 2 diabetes, especially in a dominant model (OR 1.26 [95% CI 1.03–1.53]) (Table 2). Adjustment for covariates strengthened the association. When multiple testing was controlled by permutation test, in which case-control status was randomly permuted 10,000 times for the polymorphisms examined, the association between rs1139646 became marginally significant (P = 0.08). Polymorphisms rs10920531 and rs1342387 also showed normally significant associations with diabetes risk in the multivariate analyses. However, these associations did not remain significant after adjustment for multiple testing.

In the haplotype analyses, five common haplotypes (frequency >0.05) possessing the tagging SNPs of ADIPOR1 were identified that accounted for 84% allelic variance. ADIPOR1 haplotypes showed significant associations with diabetes risk (overall test, −2log-likelihood = 15.1 on 5 df, P = 0.0098) (Table 3). A single copy of haplotype 001100 was associated with 24% decreased diabetes risk compared with the most common haplotype 110000 (OR 0.76 [95% CI 0.61–0.96], P = 0.02). Figure 2 presents a cladogram representing the inferred evolutionary relatedness of the haplotypes. Thus, no single polymorphism could distinguish haplotype 001100 from the reference haplotypes to suggest a functional variation.

None of the 16 ADIPOR2 polymorphisms was significantly associated with diabetes risk. By using all the tagging polymorphisms, we inferred five common haplotypes (>5%) that account for 64.4% allelic variance. These haplotypes were not significantly associated with diabetes risk (data not shown). Haplotypes inferred from polymorphisms within each LD block (Fig. 1) were not significantly associated with the disease either. We further extensively examined the haplotype associations using a sliding window approach. Figure 3 shows the significance value [−log₁₀(P)] for the associations of haplotypes consisting of two to six markers that were fit as a sliding window across the ADIPOR2 gene with diabetes risk. None of the haplotype windows showed nominally significant associations with diabetes risk nor did the haplotype windows covering more polymorphisms (more than six SNPs).

DISCUSSION

In this prospective, nested, case-control study of 714 incident cases of type 2 diabetic and 1,120 matching control female subjects, we found that the haplotypes inferred from the LD tagging SNPs of the ADIPOR1 gene were significantly associated with the risk of type 2 diabetes. ADIPOR2 polymorphisms were not associated with the risk of type 2 diabetes.

Adiponectin is one of the most abundant adipose tissue–specific cytokines and has shown protective effects against insulin resistance, type 2 diabetes, and cardiovascular disease (26). It has been documented that adiponectin might increase insulin sensitivity by enhancing insulin's suppressive effect on gluconeogenesis and by promoting glucose utilization and fatty acid oxidation in liver and skeletal muscle (27,28). Recently, two isoforms of the cellular receptor for adiponectin, ADIPOR1 and ADIPOR2, were cloned and characterized. The metabolic and insulin-sensitizing effects of adiponectin are believed to be mediated by these receptors via activation of adenosine monophosphate–activated protein kinase and increasing peroxisome proliferator–activated receptor-α ligand activities (7,29). In mice, ADIPOR1 is ubiquitously expressed, whereas ADIPOR2 expression is more restricted to skeletal muscle and liver. Unlike in mice, both receptors are found to be highly expressed in skeletal muscle (30) and in pancreatic β-cells in humans (31). ADIPOR1 is a high-affinity receptor for globular adiponectin, while ADIPOR2 is an intermediate-affinity receptor for both globular and full-length adiponectin (7,29). Expression of ADIPOR1 and ADIPOR2 has been associated with insulin resistance (32) and was found to be lower in diabetic and normal glucose-tolerant subjects with family history of diabetes (10,30).

Several studies have documented that the variations in the ADIPOR1 gene were related with adiposity, insulin resistance (11–13), and the risk of type 2 diabetes (14–17). Our findings that haplotypes possessing the tagging SNPs of the ADIPOR1 gene were associated with diabetes risk lend further credibility for the potential roles of the variability of the ADIPOR1 gene in the etiology of the disease. Earlier evidence indicates that the 3′ UTR polymorphism rs1139646 may affect the expression of the ADIPOR1 gene (15). Although rs1139646 showed the strongest and nominally significantly association with diabetes risk in the present study, it is less likely to be a causal variant because such an association was substantially attenuated after control of multiple testing. We assume the causal variants may be cosegregated with the risk haplotype identified. It is also possible that the observed associations were due to the effects of specific allele combination of several variations, which likely occur in the haplotype-driven selection (33).

Different haplotypes of the ADIPOR2 gene were associated with the type 2 diabetes or combined type 2 diabetes/impaired glucose tolerance trait in two studies of Amish and French (14,17) but not in Asian Japanese (16) subjects. Of note, there is not a single genetic marker that has shown repeatable effect. We did not find significant associations between the variations in this gene, individually or in haplotypes, with diabetes risk in U.S. women. We have genotyped several previously reported polymorphisms including rs11061971, rs1044471, rs1342387 (14), and rs1139646 (15). Some other reported polymorphisms, rs767870, rs2286380 (tagged by rs7316374) (17), rs12342 (tagged by rs11061937), and rs2275737 (tagged by rs1342387) (14), could be well captured by the tagging SNPs used in our study (with r² > 0.8).

The cause of discrepancy among the studies is unknown but may be partly due to the diversity in ethnicity and population compositions, especially the sex profiles. In the present study, the genetic associations were assessed in women, while most previous studies included both men and women (14–17). The sex difference in the phenotypes of obesity and type 2 diabetes is well known. It has also been documented that adiponectin concentrations were higher in women than in men (34). A recent study suggests that the associations between circulating adiponectin levels and diabetes risk may be different in men and women (35). In addition, body fat may affect the expression of adiponectin and ADIPORs (29,36). The polymorphisms of ADIPORs have been inconsistently related to adiposity measurements in some studies (12,13,15). However, we did not find significant genetic associations with BMI in U.S. women. Again, we suspect the discrepant observations between various studies may be partly attributed to the difference in population compositions. Further adjustment for BMI did not materially change the associations between ADIPOR polymorphisms and diabetes risk.

Limited evidence suggests that the genetic variability of ADIPOR1 may be more relevant to insulin resistance compared with ADIPOR2 (12). ADIPOR1 mRNA expression in human skeletal muscle cells was positively correlated with in vivo insulin secretion and plasma lipid concentrations, whereas the expression of ADIPOR2 was not associated with most of these metabolic parameters (9). In addition, it was recently found that regulation of ADIPOR1, rather than ADIPOR2, might be involved in glucose and lipid metabolism in diabetic states (37).

The major strengths of this study include the prospective nature in design and large sample size. Most previous studies examined the associations using prevalent cases. The patient selection–inherent bias may skew the associations because the prevalence of diabetes is likely to be affected by life expectancy, which is subjected to the genetic effects on other conditions. In the analyses, we carefully controlled the potential influence of multiple comparisons using permutation tests. Attempts to address this issue can reduce the probability of false-positive results and help guide the interpretation of the results. As a limitation, population stratification may influence the observed associations. However, our populations are racially homogeneous, with the majority of the participants being white (∼96%). Further adjustment for ethnicity or removing the minorities from the analyses did not appreciably change the results. In addition, our analyses were restricted to women and therefore may be not generalizable to men.

In conclusion, we found that the genetic variability in ADIPOR1 was associated with the risk of type 2 diabetes. Our data did not support substantial associations between ADIPOR2 variability and diabetes risk. Further analyses are warranted to replicate the associations in other populations, to identify the potential functional variants, and to elucidate the metabolic changes that may affect diabetes risk.