Uncoupling Protein 2 Promoter Polymorphism −866G/A, Central Adiposity, and Metabolic Syndrome in Asians

Authors

  • Haiqing Shen,

    1. Nutrition and Genomics Laboratory, U.S. Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston, Massachusetts
    Search for more papers by this author
  • Lu Qi,

    1. Nutrition and Genomics Laboratory, U.S. Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston, Massachusetts
    Search for more papers by this author
  • E. Shyong Tai,

    1. Department of Endocrinology, Singapore General Hospital, Singapore
    Search for more papers by this author
  • Suok Kai Chew,

    1. Department of Endocrinology, Singapore General Hospital, Singapore
    Search for more papers by this author
  • Chee Eng Tan,

    1. Epidemiology and Disease Control Division, Ministry of Health, College of Medicine Building, Singapore.
    Search for more papers by this author
  • Jose M. Ordovas

    Corresponding author
    1. Nutrition and Genomics Laboratory, U.S. Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, Boston, Massachusetts
      Nutrition and Genomics Laboratory, U.S. Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, 711 Washington Street, Boston, MA 02111. E-mail: jose.ordovas@tufts.edu
    Search for more papers by this author

  • 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.

Nutrition and Genomics Laboratory, U.S. Department of Agriculture Human Nutrition Research Center on Aging, Tufts University, 711 Washington Street, Boston, MA 02111. E-mail: jose.ordovas@tufts.edu

Abstract

A polymorphism in the promoter region of uncoupling protein 2 gene −866G/A has been associated with its expression levels in adipose tissue, the risk of obesity, and metabolic abnormalities. Our purpose was to examine the associations of −866G/A with body fat and the risk of metabolic syndrome in a random sample of 4018 Asians (1858 men and 2160 women) from three ethnic groups (Chinese, Malay, and Indian). The minor allele frequency of −866G/A polymorphism in South Asians was similar to that in whites. After adjustment for covariates including age, cigarette smoking, and physical activity, the −866A/A genotype was associated with higher waist-to-hip ratio as compared with the wild-type genotype in Chinese and Indian men (p = 0.018 and p = 0.046, respectively). Moreover, Indian men with −866A/A genotype had a significantly increased risk of metabolic syndrome as compared with those homozygous for the wild-type (odds ratio, 2.66; 95% confidence interval, 1.21 to 5.88; p = 0.015). Such a risk was mainly caused by the excess presence of hypertriglyceridemia and central obesity. Our findings indicate that the uncoupling protein 2 gene −866G/A polymorphism may increase the risks of central obesity and metabolic syndrome, with greater effects on Asian men.

Uncoupling protein 2 (UCP2)1 is a member of the mitochondrial transporter superfamily that uncouples oxidative phosphorylation from adenosine triphosphate synthesis and releases excess energy as heat (1). UCP2 is expressed in multiple tissues, with predominant expression in white adipose tissue and skeletal muscle (1). In addition to the uncoupling activity seen in yeast and mammalian cells (1, 2), UCP2 may modulate insulin action (3) and lipid metabolism (4). In humans, the expression of UCP2 mRNA in adipose tissue is lower in the obese subjects as compared with the lean controls, suggesting UCP2 may be involved in the etiology of obesity (5).

The human UCP2 gene is located on chromosome 11q13. A common polymorphism at the UCP2 promoter region, −866G/A, has been associated with increased mRNA expression of the UCP2 gene in adipose tissue and with an enhanced in vitro transcription (6). The association between the −866G/A polymorphism and obesity was observed in the initial study (6), although it was not observed in the following studies (7, 8). In addition, the −866G/A polymorphism has also been associated with several components of metabolic syndrome such as dyslipidemia (8), insulin resistance (9), hypertension (7), and type 2 diabetes (9, 10, 11). However, such associations have yet to be replicated in different populations.

In this study, we investigated the associations of the UCP2 −866G/A polymorphism with body fatness and the risk of metabolic syndrome in a large and well-characterized multiethnic Asian population.

Among 4018 unrelated study subjects (68.1% Chinese, 18.3% Malays, and 13.6% Indians) participating in the 1998 Singapore National Health Survey (NHS 98), the allele frequency of UCP2 −866A was similar across ethnic groups (0.38, 0.38, and 0.39 for Chinese, Malays, and Indians, respectively; p = 0.89) and comparable with those reported in whites (8, 9, 10). The genotype distribution did not deviate from the Hardy-Weinberg equilibrium (p > 0.05). Table 1 shows some relevant characteristics of the participants according to ethnic group and gender. Indians had the highest prevalence of metabolic syndrome for both men and women.

Table 1.  Descriptive characteristics of the study subjects according to ethnic and gender groups in the Singapore NHS 98 population
 Men
 Chinese (n = 1238)Malay (n = 355)Indian (n = 265)p
  1. Continuous variables are presented as mean ± standard deviation (shown in parentheses), whereas categorical variables are presented as percentages of prevalence. p Value obtained in the comparison among three genotype groups (ANOVA test for means and χ2 test for percentages). NHS 98, 1998 Singapore National Health Survey; HDL-C, high-density lipoprotein cholesterol; TG, triglyceride.

Age (years)38.0 (12.2)39.6 (12.6)41.2 (12.1)0.0003
BMI (kg/m2)23.5 (3.7)24.7 (4.1)24.7 (4.4)<0.0001
Waist (cm)84.1 (9.8)85.7 (10.5)89.4 (11.3)<0.0001
Fasting glucose (mM)5.79 (1.35)6.13 (2.00)6.39 (2.37)<0.0001
HDL-C (mM)1.26 (0.32)1.15 (0.28)1.06 (0.28)<0.0001
TGs (mM)1.70 (1.70)2.00 (1.60)2.09 (1.81)<0.0001
Systolic blood pressure (mm Hg)125.2 (14.8)125.5 (17.7)124.3 (16.3)0.63
Diastolic blood pressure (mm Hg)77.4 (10.6)77.7 (11.9)77.7 (11.5)0.91
Metabolic syndrome (%)22.131.040.4<0.0001
Diabetes status (%)7.011.018.9<0.0001
 Women
 Chinese (n = 1498)Malay (n = 381)Indian (n = 281)p
Age (years)37.7 (12.2)38.6 (12.9)40.0 (12.0)0.01
BMI (kg/m2)22.0 (3.6)26.3 (5.6)25.5 (5.1)<0.0001
waist (cm)73.0 (8.4)79.5 (12.4)81.5 (11.5)<0.0001
Fasting glucose (mM)5.49 (1.26)6.20 (2.60)6.13 (2.05)<0.0001
HDL-C (mM)1.56 (0.36)1.44 (0.33)1.23 (0.30)<0.0001
TGs (mM)1.16 (0.75)1.40 (0.88)1.35 (0.69)<0.0001
Systolic blood pressure (mm Hg)117.1 (16.6)124.5 (21.3)117.8 (17.6)<0.0001
Diastolic blood pressure (mm Hg)70.7 (10.9)74.8 (12.3)69.4 (11.2)<0.0001
Metabolic syndrome (%)12.329.130.6<0.0001
Diabetes status (%)6.415.217.9<0.0001

The degree of obesity and fat distribution differed across genders and ethnic groups. Thus, we analyzed the associations separately according to ethnic and gender groups to avoid the potential confounding effects caused by genetic and environmental influences related to ethnicity and gender. We found that the −866A/A genotype was associated with higher waist-to-hip ratio as compared with the wild-type genotype −866G/G in Chinese and Indian men (p = 0.018 and p = 0.046, respectively). Such associations were independent of age, cigarette smoking, and physical activity. The genotypic difference (A/A vs. G/G) in waist was 2.1 cm for Chinese (p = 0.016). Moreover, subjects with the A/A genotype had trends toward increased body weight in Chinese men and increased BMI in Indian men, although these were not statistically significant (p = 0.055 and p = 0.05, respectively) (Table 2). When all ethnic groups were analyzed together, the observed associations in men (p = 0.047 for waist and p = 0.032 for waist-to-hip ratio) had the same trends as those found in the individual ethnic groups. The UCP2 −866G/A polymorphism was not significantly associated with anthropometric measures in Malay men (Table 2) and Asian women (data not shown).

Table 2.  Effect of UCP2 −866G/A polymorphism on traits related to body fatness in men from each ethnic group
 Chinese men Malay men Indian men 
 −866G/G (n = 467)−866G/A (n = 573)−866A/A (n = 198)p−866G/G (n = 125)−866G/A (n = 173)−866A/A (n = 57)p−866G/G (n = 94)−866G/A (n = 130)−866A/A (n = 41)p
  • All of the obesity-related measures were adjusted for age, cigarette smoking, and physical activity (presented by means and standard error, shown in parentheses). p Value based on multiple linear regression. UCP2, uncoupling protein 2.

  • *

    p< 0.05 vs. genotype groups G/G.

  • p< 0.05 vs. genotype groups G/A.

Age (years)37.9 (0.6)38.6 (0.5)36.8 (0.9)0.17439.8 (1.1)40.4 (1.0)36.8 (1.7)0.17441.8 (1.3)40.7 (1.1)41.4 (1.9)0.770
Weight (kg)67.2 (0.5)67.4 (0.5)69.4 (0.8)0.05569.3 (1.1)69.0 (0.9)66.2 (1.6)0.23969.5 (1.3)71.9 (1.1)74.2 (2.0)0.129
BMI (kg/m2)23.4 (0.2)23.5 (0.2)24.0 (0.3)0.15924.9 (0.4)24.8 (0.3)23.7 (0.5)0.14324.0 (0.4)24.9 (0.4)25.9 (0.7)0.050
Waist (cm)83.7 (0.4)83.8 (0.4)85.8 (0.7)*,0.01686.6 (0.9)85.8 (0.7)83.7 (1.3)0.18487.6 (1.1)89.9 (1.0)92.1 (1.7)0.071
Hip (cm)95.8 (0.3)95.8 (0.3)96.8 (0.5)0.14898.0 (0.7)98.2 (0.6)96.5 (1.0)0.32798.1 (0.9)99.3 (0.8)100.5 (1.3)0.291
Waist-to-hip ratio0.87 (0.00)0.87 (0.00)0.88 (0.00)*,0.0180.88 (0.00)0.87 (0.00)0.87 (0.01)0.1290.89 (0.01)0.90 (0.01)0.91 (0.01)*0.046

Next, we evaluated the association between the UCP2 −866G/A and the risk of metabolic syndrome. Among Indian men, the prevalence of metabolic syndrome was significantly different across genotypes, with higher proportion in individuals with A/A genotype (24 of 41, 58.5%) and lower proportion in heterozygotes G/A (48 of 130, 36.9%) and the wild-type homozygotes G/G (35 of 94, 37.2%), (p = 0.036). Using logistic analysis, the A/A genotype was associated with ∼2.6 times higher risk of metabolic syndrome as compared with the G/G genotype [odds ratio (OR), 2.66; 95% confidence interval (CI), 1.21 to 5.88; p = 0.015]. Adjustment for age, cigarette smoking, and physical activity did not appreciably change the association (OR, 2.67; 95% CI, 1.20 to 5.91; p = 0.016). G/A subjects did not have an increased risk of metabolic syndrome as compared with the G/G genotype (OR, 1.07; 95% CI, 0.60 to 1.92; p = 0.82). When the individual components of metabolic syndrome were analyzed, we found that abdominal obesity and hypertriglycemia were more prevalent in the A/A genotype than in the G/G genotype (58.54% vs. 38.30%, p < 0.05 for abdominal obesity; 73.17% vs. 43.62%, p < 0.01 for hypertriglycemia) (Figure 1). We also analyzed the relation between the UCP2 −866G/A polymorphism and individual components of metabolic syndrome as continuous variables and found similar results to those using categorical variables. Besides the results for abdominal obesity reported in Table 2, our analyses showed that triglyceride (TG) levels were significantly elevated in Indian men with the UCP2 −866A/A genotype (2.93 ± 0.27 mmol/dL) relative to −866G/A (1.97 ± 0.15 mmol/dL) or −866G/G genotype (1.89 ± 0.18 mmol/dL) (p = 0.026 A/A vs. G/A, p = 0.018 A/A vs. G/G, respectively). We did not find significant association between the UCP2 −866G/A and other metabolic syndrome-related traits in Indian men. In addition, −866G/A was not associated with metabolic syndrome in other ethnic or gender groups.

Figure 1.

The prevalence of metabolic traits according to the UCP2 −866G/A genotype in Asian Indian men. p is the significance of the trends across the genotype groups by χ2 test. * or † represent p < 0.01 by χ2 test for comparison between genotype groups G/G and A/A or G/A and A/A, respectively. ‡ p < 0.05 by χ2 test for comparison between genotype groups G/G and A/A. Ab, abdominal; Glu, glucose; BP, blood pressure.

In this cross-sectional study of 4018 Asians from three ethnic groups (Chinese, Indian, and Malay), we found that the UCP2 −866G/A genotypes were significantly associated with two indicators of central obesity, i.e., waist-to-hip ratio and waist circumference among Chinese and Indian men. Moreover, Indian men carrying the −866A/A genotype were at higher risk of metabolic syndrome. Such risk was mainly caused by the excess presence of hypertriglyceridemia and central obesity among those individuals.

In this study, the UCP2 −866A/A genotype allele of UCP2 was associated with increased central fat in Chinese and Indian men. Such an association was in line with recently reported functional effects of −866G/A on UCP2 expression, by which −866A/A subjects had decreased UCP2 mRNA levels in adipose tissue relative to −866G/G subjects (12). The lower expression at the transcription level may result in a decreased production of UCP2, decreased energy expenditure, and, hence, increased accumulation of body fat. Although −866G/A was associated with a decreased risk of obesity in an initial study (6), such an association was not replicated by others (7, 8, 13). In fact, one study on Danish people found a modest association with an increased BMI in young and lean men (13). Moreover, in another study of diabetic and non-diabetic individuals from families with type 2 diabetes, significant associations of UCP2 haplotypes were observed with BMI, and the highest values were found in individuals having a haplotype that was heterozygous at the −866G/A locus (12). The discrepancy of studies may be due to the modest effect of this polymorphism, which is more likely to be modified by environmental or lifestyle confounders. The genetic heterogeneity of human obesity itself may also account for the diverse associations. The statistical significance of our results was restricted to Chinese and Indians. However, we cannot conclude that the observed associations are race specific. Given the small effects associated with this UCP2 polymorphism, a larger sample size is needed for Malays.

In addition to its relation with central adiposity, variant −866G/A was also associated with significantly increased risk of metabolic syndrome in Indian men, and such an association appears to be driven by the excess presence of hypertriglyceridemia and central adiposity. It is noteworthy that Indians showed the highest prevalence of metabolic syndrome among the ethnicities examined. Our findings agree with previously reported associations between −866G/A and the components of metabolic syndrome including hypertriglyceridemia (8), insulin resistance (9), hypertension (7), and type 2 diabetes (10). Also, it was recently reported that hypertriglyceridemia (≥1.7 mM) was more frequent in −866A/A homozygous carriers than in G/G homozygotes among diabetic patients (8).

Several lines of evidence support the roles of UCP2 in the homeostasis of TGs (3, 4). UCP2 may participate in the regulation of lipids as fuel substrates, possibly through its function as free fatty acid transporter to facilitate high rates of lipid oxidation in mitochondrial membrane (4). In addition, UCP2 may play an indirect role in modulating free fatty acid metabolism by regulating β-cell insulin secretion (3). In addition, UCP2 could also influence plasma TG levels by its effects on central obesity, which is more metabolically active and sensitive to lipolysis (14).

We did not find relevant associations of −866G/A in women. This may be due to their lower visceral adiposity. Additionally, sex hormones have been implicated in differential expression of UCP2 (15, 16). Therefore, it should be interesting to investigate in further studies whether hormonal status (i.e., estrogens) attenuates the association of this variant with metabolic and anthropometric traits.

UCP2 −866G/A has been associated with type 2 diabetes in several studies (9, 10, 11). Recently, Bulotta et al. (17) have found that the minor allele of UCP2 −866G/A was associated with decreased risk of type 2 diabetes in whites. The finding contradicts previous results in whites and Asians (9, 10, 11). Although we found that the genetic variants correlated with significant increased risk of metabolic syndrome in Indian men, we did not find a significant association with type 2 diabetes. A much larger sample size or another study design, e.g., case-control study, may help draw definitive conclusion about the relation with type 2 diabetes in Asians. Moreover, because of the limited statistical power, especially in the subethnic groups, we cannot exclude the possibility of a false-positive association.

In summary, we found significant associations among the UCP2 −866G/A polymorphism, central adiposity, and increased risk of metabolic syndrome in Asian men. Our findings lend support to the roles of UCP2 in the development of obesity and related metabolic disorders.

Research Methods and Procedures

Subjects and Methods

In total, 4018 subjects (1238 Chinese, 355 Malay, and 265 Indian men and 1498 Chinese, 381 Malay, and 281 Indian women) who participated in the NHS 98 were included in this study. NHS 98 was an initiative to determine the risk factors for the major non-communicable diseases in Singapore. The detailed methodology has been described elsewhere (18). In brief, based on the protocols and procedures recommended by the World Health Organization for field surveys of diabetes and other non-communicable diseases and the World Health Organization Multinational Monitoring of Trends and Determinants in Cardiovascular Disease protocol for population survey, NHS 98 conducted systematic sampling according to household types, followed by disproportionate stratified sampling by ethnic groups. Oversampling was applied for Malays and Asian Indians. Informed consent was obtained from all participants in the survey. The study was approved by the Ministry of Health, Singapore. Data on lifestyle factors and biochemistry parameters were collected as previously described (19).

Metabolic syndrome was defined using the modified National Cholesterol Education Program Adult Treatment Panel III guidelines. It is characterized by the presence of three or more of the following characteristics: elevated TGs (>1.7 mM), low high-density lipoprotein cholesterol (HDL-C; <1.0 mM for men and <1.3 mM for women), high blood pressure (≥130/85 mm Hg or known treatment for hypertension), elevated fasting glucose (>6.1 mM or known treatment for diabetes), and abdominal obesity (in Asian populations using reduced values for abdominal waist circumference, >90 cm for men and >80 cm for women) (20). We considered subjects being treated for hypertension as having blood pressure ≥ 130/85 mm Hg. Similarly, subjects being treated for diabetes, regardless of the fasting glucose level, were considered as having fasting glucose > 6.1 mM (20).

DNA extraction was carried out using QIAamp DNA blood Midi kits (Qiagen, Hilden, Germany) according to the manufacturer's recommended protocol. Genotyping was carried out using ABI Prism SnapShot multiplex system (Applied Biosystems, Foster City, CA). The 383-base pair DNA fragment encompassing the single nucleotide polymorphism was amplified by polymerase chain reaction using the following primers: 5′- GCAAGATCTCCTCATGGCAGAAATA-3′ and 5′- CTTTAATTGGCTGACCCGTCCTGT-3′. The single nucleotide polymorphism-specific probe 5′-CTTTGTGGCCTACCAA-3′ was used in the single-nucleotide extension process.

Statistical Analysis

χ2 Tests were used to test the Hardy-Weinberg equilibrium of the genotypes and to compare the distribution of genotypes between genders and between the ethnic groups. The influence of covariates in the comparison of means among genotype groups was controlled by multiple linear regression analysis. Log transformation of TG and fasting glucose was performed to improve normality of the distribution. Tukey's procedure was used for multiple comparisons between groups. The major covariates included age, cigarette smoking, and physical activity. χ2 Tests were used to detect the difference between number of people with metabolic syndrome and without metabolic syndrome across the genotypes. Multivariate logistic regression analysis was used to control for covariates and estimate the ORs. SAS (Windows version 8.0; SAS Institute, Cary, NC) was used to analyze the data, and statistical significance was defined at the 5% nominal level.

Acknowledgement

This work was supported by NIH/National Heart, Lung, and Blood Institute Grant HL54776, by U.S. Department of Agriculture Research Service Contracts 53-K06-5-10 and 58-1950-9-001, and by National Medical Research Council of Singapore Grant 0462/2000.

Footnotes

  • 1

    Nonstandard abbreviations: UCP2, uncoupling protein 2; NHS 98, 1998 Singapore National Health Survey; OR, odds ratio; CI, confidence interval; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol.

Ancillary