Introduction

It has been found that body fat distribution is an important factor in coronary heart disease (CHD). In particular, a large waist circumference or waist-to-hip ratio (WHR) is closely associated with an increased prevalence of risk factors for CHD, for example, impaired glucose tolerance, insulin resistance, lipoprotein metabolic disorder and hypertension.^1,2,3 Though, in general, upper-body or android-type⁴ obesity, with a large waist or WHR, is more frequently observed in men compared with women, this obesity phenotype is also observed fairly often in postmenopausal women.⁵ This is because estrogen deficiency during the normal menopausal transition accelerates the selective deposition of intra-abdominal fat.⁶ With respect to estrogen's association to body fat distribution, several authors ^6,7,8,9,10 have reported that estrogen hormone replacement therapy had desirable effects on body fat distribution in postmenopausal women. These findings suggest that estrogen plays an important role in the modification of body fat distribution.

More recently, associations have been found between estrogen receptor (ER) gene polymorphism and bone mineral density,^{11,12,13,14,15,16,17,18} pathogenesis of type II diabetes,¹⁹ and susceptibility to or age of onset of autoimmune diseases such as multiple sclerosis.²⁰ The human ER gene is located on chromosome 6p25.1, is comprised of eight exons, and spans >140 kb.²¹ Two single-nucleotide polymorphisms (SNPs) have been identified in the first intron of the ER gene: a T C polymorphism that is recognized by the restriction endonuclease PvuII (T and C alleles correspond to the presence (p allele) and absence (P allele) of the restriction site, respectively) and an A G polymorphism that is recognized by XbaI (A and G alleles correspond to the presence (x allele) and absence (X allele) of the restriction site, respectively). According to studies on the relation between the ER gene polymorphism and bone mineral density, the SNPs, alone or in combination, were associated with lower bone mineral density in pre- and postmenopausal women,^{11,14,15,17,18,22,23} and were important factors in determining changes in bone mass in older women receiving hormone replacement therapy.^16,24

To our knowledge, little has been reported on the association between the ER gene polymorphisms and body fat distribution. In clinical settings for obesity treatment, understanding this association would be helpful not only for early preventative treatment of upper-body obesity but also for predicting the effects of estrogen replacement therapy on the modification of body fat distribution. The purpose of this study, therefore, was to examine whether the T C (PvuII) and A G (XbaI) polymorphisms of the ER gene, alone or in combination, are associated with body fat distribution in a middle-aged to elderly Japanese population.

Methods

Subjects

There were 1110 women and 1128 men who participated in the first wave of examinations in the National Institute for Longevity Sciences–Longitudinal Study of Aging (NILS-LSA) from April 1998 to March 2000. There were randomly sampled, community-dwelling individuals aged 40–79 y, stratified by age and gender and living in the neighborhood of the NILS. Details of the NILS-LSA have been described elsewhere.²⁵ The aim and design of the study were explained to each subject before they gave their written informed consent. The study was approved by the Committee of the Chubu National Hospital.

Determination of ER genotypes

The ER genotypes were determined in accordance with a study by Yamada et al.²⁶. The ER gene was analyzed with an automated fluorescent allele-specific DNA primer assay system (Toyobo Gene Analysis, Osaka, Japan). To determine the T C (PvuII) genotype, the polymorphic region of the gene was amplified by polymerase chain reaction with allele-specific sense primers labeled at the 5' end either with fluorescein isothiocyanate (5'-AGTTCCAAATGTCCCAGXTG-3') or with Texas red (5'-AGTTCCAAATGTCCCAGXCG-3') and an antisense primer labeled at the 5' end with biotin (5'-TCTGGGAAACAGAGACAAAGC-3'). The reaction mixture (25 l) contained 20 ng of DNA, 5 pmol of each primer, 0.2 mmol/l of each deoxynucleoside triphosphate, 2.5 mmol/l MgCl₂, and 1 U of DNA polymerase (rTaq; Toyobo) in rTaq buffer. The amplification protocol consisted of three parts: initial denaturation at 95°C for 5 min; 35 cycles of denaturation at 95°C for 30 sec, annealing at 62.5°C for 30 sec, and extension at 72°C for 30 sec; and a final extension at 72°C for 2 min.

To determine the A G (XbaI) genotype, the polymorphic region of the gene was amplified by polymerase chain reaction with a sense primer labeled at the 5' end with biotin (5'-CTGTTTCCCAGAGACCCTGAG-3') and allele-specific antisense primers labeled at the 5' end either with fluorescein isothiocyanate (5'-CCAATGCTCATCCCAACTXTA-3') or with Texas red (5'-CCAATGCTCATCCCAACTXCA-3'). The reaction mixture (with the exception of the primers) and the amplification protocol (with the exception that the annealing temperature was 65°C) were identical to those used for genotyping the T C (PvuII) polymorphism.

Amplified DNA was incubated in a solution containing streptavidin-conjugated magnetic beads in the wells of a 96-well plate at room temperature. The plate was placed on a magnetic stand, and the supernatants were then collected from each well, transferred to the wells of a 96-well plate containing 0.01 M NaOH, and measured for fluorescence with a microplate reader (Fluoroscan Ascent; Dainippon Pharmaceutical, Osaka, Japan) at excitation and emission wavelengths of 485 and 538 nm, respectively, for fluorescein isothiocyanate and at 584 and 612 nm, respectively, for Texas red.

The T C and A G polymorphisms were determined in 2228 subjects (1108 women, 1120 men) and in 2235 subjects (1107 women, 1128 men), respectively.

Anthropometric variables

Body weight was measured to the nearest 0.01 kg. using a digital scale, height was measured to the nearest 0.1 cm using a wall-mounted stadiometer, and body mass index (BMI) was calculated as weight (kg) divided by height squared (m²). Waist circumference and WHR were used as the indices for body fat distribution in this study. WHR was calculated as a ratio of waist circumference measured at the level of the umbilicus to hip circumference.

Body composition by dual-energy X-ray absorptiometry

Whole-body fat mass (FM), fat-free mass (FFM) and percentage FM (%FM), assessed by dual-energy X-ray absorptiometry (QDR-4500; Hologic, Madison, OH, USA), were used as the indices for determining body composition. Transverse scans were used to measure FM and FFM, and pixels of soft tissue were used to calculate the ratio (R value) of mass attenuation coefficients at 40–50 keV (low energy) and 80–100 keV (high energy), using software version 1.3Z.

Physical activity, smoking status, menstrual status and hormonal replacement therapy

A detailed interview with questionnaire sheets ^27,28 revealed work time and leisure time physical activities of the subjects. Amount of physical activity was calculated as a product of the metabolic-equivalent (MET) by duration in minutes. Smoking status, menstrual status and hormonal replacement therapy were examined by a medical doctor. Menopause was defined as the absence of menses for at least 12 months by a questionnaire.

Biochemical assays of blood

An antecubital blood sample was drawn from each subject after an overnight fast. Serum total cholesterol and triglycerides were determined enzymatically, serum high-density lipoprotein cholesterol was measured by the heparin-manganese precipitation method and fasting plasma glucose was assayed by a glucose oxidase method. Plasma insulin was measured in duplicate by radioimmunoassay. Serum low-density lipoprotein cholesterol was estimated according to the Friedewald formula.²⁹

Data analysis

To examine whether age may influence the relation between the ER gene polymorphisms and body fat distribution, we subdivided each gender group into two age groups: middle-aged (40–59 y) and older (60–79 y). Furthermore, middle-aged women were categorized by menopause status. Values are expressed as meanstandard error (s.e.) in the tables and figures. Allele frequencies were estimated by the gene-counting method, and the ² test was used to identify significant departures from Hardy–Weinberg equilibrium. The distribution of haplotypes for the T C and A G polymorphisms was calculated according to the method by Thompson et al.³⁰ The data were compared by one-way analysis of variance and the Tukey–Kramer post hoc test. When a significant difference exists, analysis of covariance was used with age, smoking status, menstrual status, hormonal replacement therapy and physical activities as covariates. The relations between age and both FM and waist were tested by correlation analysis. In each statistical analysis, probability values below 0.05 were regarded as significant. The data were analyzed with the Statistical Analysis System (SAS), release 6.12.

Results

Physical and biochemical blood characteristics of the subjects are shown for each gender in Table 1. No difference was found in BMI, waist, hip or WHR between genders, whereas %FM and FM were significantly greater in women than in men. The distributions of ER genotypes with regard to the T C and A G SNPs were in Hardy–Weinberg equilibrium for the subjects (see Table 2). There was no difference in the distribution between genders or between age groups. The distribution of haplotypes for the T C and A G polymorphisms in all study subjects was as follows: T/A, 61.9%; T/G, 0.2%; C/A, 28.6%; and C/G, 9.3%. The T C and A G SNPs were in linkage disequilibrium (pairwise linkage disequilibrium coefficient, D' (D/D_max), of 0.97; standardized linkage disequilibrium coefficient, r, of 0.40; P<0.0001, ² test).

Table 1 - Physical and biochemical blood characteristics of subjects (n=2238).

Full table

Table 2 - Distribution of T C(Pvu II) and A G(Xba I) genotypes of the ER gene of the subjects.

Full table

T C polymorphisms (PvuII)

For both genders, no significant difference was found in any variables among the T C polymorphisms with the exception of BMI of the older men (table not shown). Meanss.e. of BMI of the older men were 22.90.2, 22.50.2 and 22.00.3 kg/m² in the TT, TC and CC genotypes, respectively (TT>CC, P=0.025).

A G polymorphism (XbaI)

For middle-aged women with the AG and/or GG genotypes, BMI, %FM, FM, waist, hip and WHR were greater than in the middle-aged women with the AA genotype (Table 3). In particular, those with the GG genotype had a 9% greater BMI, a 20% greater FM and a 9% larger waist compared with the AA genotype. Fasting insulin was significantly higher in the individuals with the AG genotype (8.30.3 U/ml), compared with the AA genotype (7.40.2 U/ml) (see Figure 1). No difference was found in plasma lipids and fasting blood glucose among the genotypes. When the analysis of covariance with age, smoking status, menstrual status, hormonal replacement therapy and physical activities as covariates was used, these results remained essentially unchanged with the exception of hip circumference and WHR (P<0.1). Significant differences (P<0.05) were still observed in BMI (AA 22.30.3, AG 23.70.4, GG 24.41.0: AG and GG > AA), %FM (AA 29.90.4, AG 31.90.6, GG 33.11.5: AG and GG > AA), FM (AA 16.00.4, AG 18.20.6, GG 19.11.5: AG and GG > AA), waist (AA 81.30.7, AG 85.51.1, GG 88.22.5: AG and GG > AA) and fasting insulin (AA 7.30.2, AG 8.40.4, GG 8.00.9: AG > AA) among the genotypes. For older women with the GG genotype, %FM, FM and waist were smaller by 10, 18 and 6%, respectively, compared to the older women with the AA and/or AG genotypes. These results were also unchanged when the analysis of covariance was used. Significant differences (P<0.05) were still observed in %FM (AA 32.10.5, AG 33.70.7, GG 27.51.8: AA and AG > GG) and FM (AA 16.40.4, AG 18.10.6, GG 13.41.6: AA and AG > GG) among the genotypes. No difference was found in any variables among the genotypes in men (table was not shown).

Figure 1.

Comparison of fasting insulin levels among A G genotypes in middle-aged women.

Full figure and legend (63K)

Table 3 - Physical status and Physical activities of women according to age group and A G genotype (n=1107).

Full table

Table 4 compares age, FM, waist and WHR among the A G genotypes in middle-aged women according to menstrual status. Women who could not be clearly defined as pre- or postmenopausal were excluded from this analysis. For premenopausal women with the GG genotype, FM, waist and WHR were larger than in those with the AA genotype, whereas no difference was found in postmenopausal women. These results also remained essentially unchanged with the exception of WHR (P<0.1) when the analysis of covariance with age and smoking status as covariates was used. Significant differences (P<0.05) were still observed in FM (AA 15.80.5, AG 18.00.7, GG 20.71.6: GG > AA) and waist (AA 80.20.7, AG 83.71.3, GG 88.53.2: GG > AA) among the genotypes.

Table 4 - Comparison of age, fat mass and waist among the A G genotypes in middle-aged women according to menstrual status (n=536).

Full table

Combination of the T C and A G polymorphisms

To determine whether the T C and A G polymorphisms synergistically influence body fat distribution, we compared the variables by combined genotypes (table not shown). Because of the small number of subjects, TT/AG (n=1), TT/GG (n=0) and TC/GG (n=5) were excluded from this analysis.

For middle-aged women with the CC/GG genotype, the mean values of BMI, FM and waist were significantly larger by 8%, 18% and 8%, respectively, compared to those with the TT/AA genotype. But the differences between the CC/GG and TT/AA genotypes were similar to the differences between the A G polymorphism (see Table 3). The physical activities did not differ between the CC/GG and TT/AA genotypes. For older women with the CC/GG genotype, the mean values of %FM and FM were significantly lower by 12% and 17%, respectively, compared to those with the TT/AA genotype. These results also indicate that the effects of the combined genotypes on %FM and FM are not different from the A G polymorphism alone. For both middle-aged and older men, no difference was found in any variables among the combined genotypes.

Relation of age with FM and waist

Figure 2 shows the relations between age and both FM and waist in combined data from middle-aged and older women. A significant and inverse correlation (r=-0.630, P<0.001) was found between FM and age in the group with the GG genotype. Waist circumference was positively associated with age (r=0.231, P<0.001) in the group with the AA genotype, whereas an inverse association (r=-0.504, P<0.001) was found in the group with the GG genotype.

Figure 2.

Relations of age with fat mass and waist by each A G genotype in women.

Full figure and legend (70K)

Discussion

Although we have reported in the previous study ²⁶ that the A G polymorphism of the ER gene may be associated with a greater BMI in middle-aged women, little is known about the association between the ER gene polymorphisms and body fat distribution or body composition. Rankinen et al³¹ reported in the eighth update of the human obesity gene map that 174 studies found positive associations of obesity phenotype with 58 candidate genes, but no more than one study by Speer et al¹⁹ showed any association between ER gene polymorphisms and obesity phenotype. According to the study, 29 subjects (23 women and 6 men) with android-type obesity, 69 and 31% had the AG and GG genotypes, respectively, and the AA genotype was not found.

One of the major findings of the present study is to find the association of the A G polymorphism or the combination of the T C and A G polymorphisms with not only a greater BMI but also larger %FM, FM, waist circumference and WHR in middle-aged women. Moreover, the results of Table 4 reveal that for premenopausal women, the effect of the ER gene polymorphisms on body fat distribution (FM and waist) was more significant than for postmenopausal women. These observations suggest that the greater FM of individuals with the gene mutation were due to the selective fat accumulation at the abdomen, especially the intra-abdominal cavity. Further studies on the association between the gene polymorphisms and amount of intra-abdominal fat are needed to clarify the above speculation.

An increasing fasting insulin is induced by an excess accumulation of abdominal fat.³² In addition, Cooke et al ³³ found that knocked-out ER caused adipocyte hyperplasia and hypertrophy in white adipose tissue, and is accompanied by insulin resistance and glucose intolerance in rats.³³ On the basis of the above studies, we expected that fasting insulin level in the GG genotype would be highest of all the genotypes. However, as illustrated in Figure 1, fasting insulin level in the GG genotype did not always indicate the highest values. Recently, Snijder et al ³⁴ reported that fasting insulin level was positively correlated with waist circumference but was inversely associated with hip circumference. Subjects with the GG genotype have not only larger waist but also larger hip compared with the AG genotype. Thus, their fasting insulin levels might be influenced by the two opposite functions with each other.

Regardless of the strong association between the A G polymorphism and BMI, FM, waist or WHR in middle-aged women (see Table 3), the T C polymorphism alone was not associated with any variables. Moreover, additive and synergistic effects of the genotypes were not apparent with regard to body fat distribution or body composition in the subjects. These results suggest that the A G polymorphism plays an important role in body fat distribution and body composition in middle-aged women, but the T C polymorphism does not.

Figure 2 shows that the waist circumference of women with the AA genotype increased with age, which is expected. This indicates a natural (normal) change in body fat distribution of middle-aged and older women.⁵ In contrast, results of the GG genotype revealed that FM and waist size were inversely associated with age. Consequently, our data suggest that (1) middle-aged women with the GG genotype presented with a larger FM and waist, and (2) older women with the GG genotype presented with a smaller FM and waist, compared with the AA genotype, despite the observation in the middle-aged women (see Table 3). Estrogen plays an important role in maintaining desirable fat distribution in premenopausal women. Therefore, when a functional change of the ER was induced by the gene mutation, the estrogen sensitivity is deteriorated, which possibly caused the android-type fat distribution in middle-aged and premenopausal women with the GG genotype (see Table 4). On the other hand, the smaller FM and waist of older women with the GG genotype have been possibly induced by some specific effects of the gene mutation; however, the mechanisms cannot be explained by the data from this study.

Both the T C and A G polymorphisms are found in intronic regions. Intronic changes in gene sequence may have an impact on the expression of other genes by influencing the transcription and/or stability of mRNA of those genes.^35,36 Thus, further studies on the relations of ER gene polymorphisms and body fat distribution are needed to validate the findings of this study.

Our data raise the possibility that the A G polymorphisms of the ER gene, especially the GG genotype, contribute to development of the android-type fat distribution in middle-aged and premenopausal women. In older women, this gene polymorphism may serve to decrease whole-body and abdominal fat tissue.

References

1. Kissebah AH, Peiris AN. Biology of regional body fat distribution: relationship to non-insulin-dependent diabetes mellitus. Diabetes Metab Rev 1989; 5: 83–109. | PubMed | ISI | ChemPort |
2. Kissebah AH, Vydelungum N, Murray R. Relation of body distribution to metabolic complication of obesity. J Clin Endocrinol Metab 1982; 54: 163–170.
3. Micciolo R, Bosello O, Ferrari P, Armellini F. The association of body fat location with hemodynamics and metabolic status in men and women aged 21–60. J Clin Epidemiol 1991; 6: 591–608.
4. Vague J. The degree of masculine differentiation of obesities. Am J Clin Nutr 1956; 4: 20–34. | PubMed | ISI | ChemPort |
5. Gambacciani M, Ciaponi M, Cappagli B, Benussi C, De Simone L, Genazzani AR. Climacteric modifications in body weight and fat tissue distribution. Climacteric 1999; 2: 37–44.
6. Tchernof A, Poehlman ET, Despres JP. Body fat distribution, the menopause transition, and hormone replacement therapy. Diabetes Metab 2000; 26: 12–20. | PubMed | ISI | ChemPort |
7. Davis SR, Walker KZ, Strauss BJ. Effects of estradiol with and without testosterone on body composition and relationships with lipids in postmenopausal women. Menopause 2000; 7: 395–401.
8. Perrone G, Liu Y, Capri O, Critelli C, Barillaro F, Galoppi P, Zichella L. Evaluation of the body composition and fat distribution in long-term users of hormone replacement therapy. Gynecol Obstet Invest 1999; 48: 52–55.
9. Tchernof A, Calles-Escandon J, Sites CK, Poehlman ET. Menopause, central body fatness, and insulin resistance: effects of hormone-replacement therapy. Coronary Artery Dis 1998; 9: 503–511.
10. Perry AC, Allison M, Applegate EB, Jackson ML, Miller PC. The relationship between fat distribution and coronary risk factors in sedentary postmenopausal women on and off hormone replacement therapy. Obes Res 1998; 6: 40–46.
11. Albagha OM, McGuigan FE, Reid DM, Ralston SH. Estrogen receptor alpha gene polymorphisms and bone mineral density: haplotype analysis in women from the United Kingdom. J Bone Miner Res 2001; 16: 128–134. | PubMed |
12. Becherini L, Gennari L, Masi L, Mansani R, Massart F, Morelli A, Falchetti A, Gonnelli S, Fiorelli G, Tanini A, Brandi ML. Evidence of a linkage disequilibrium between polymorphisms in the human estrogen receptor alpha gene and their relationship to bone mass variation in postmenopausal Italian women. Hum Mol Genet 2000; 9: 2043–2050. | Article | PubMed | ISI | ChemPort |
13. Ongphiphadhanakul B, Chanprasertyothin S, Payatikul P, Tung SS, Piaseu N, Chailurkit L, Chansirikarn S, Puavilai G, Rajatanavin R. Oestrogen-receptor-alpha gene polymorphism affects response in bone mineral density to oestrogen in post-menopausal women. Clin Endocrinol 2000; 52: 581–585.
14. Ogawa S, Hosoi T, Shiraki M, Orimo H, Emi M, Muramatsu M, Ouchi Y, Inoue S. Association of estrogen receptor beta gene polymorphism with bone mineral density. Biochem Biophys Res Commun 2000; 269: 537–541. | Article | PubMed | ISI | ChemPort |
15. Lorentzon M, Lorentzon R, Backstrom T, Nordstrom P. Estrogen receptor gene polymorphism, but not estradiol levels, is related to bone density in healthy adolescent boys: a cross-sectional and longitudinal study. J Clin Endocrinol Metab 1999; 84: 4597–4601. | Article | PubMed |
16. Deng HW, Li J, Li JL, Johnson M, Gong G, Davis KM, Recker RR. Change of bone mass in postmenopausal Caucasian women with and without hormone replacement therapy is associated with vitamin D receptor and estrogen receptor genotypes. Hum Genet 1998; 103: 576–585. | PubMed |
17. Kobayashi S, Inoue S, Hosoi T, Ouchi Y, Shiraki M, Orimo H. Association of bone mineral density with polymorphism of the estrogen receptor gene. J Bone Miner Res 1996; 11: 306–311. | PubMed | ISI | ChemPort |
18. Sano M, Inoue S, Hosoi T, Ouchi Y, Emi M, Shiraki M, Orimo H. Association of estrogen receptor dinucleotide repeat polymorphism with osteoporosis. Biochem Biophys Res Commun 1995; 217: 378–383.
19. Speer G, Cseh K, Winkler G, Vargha P, Braun E, Takacs I, Lakatos P. Vitamin D and estrogen receptor gene polymorphisms in type 2 diabetes mellitus and in android type obesity. Eur J Endocrinol 2001; 144: 385–389. | PubMed |
20. Niino M, Kikuchi S, Fukazawa T, Yabe I, Tashiro K. Estrogen receptor gene polymorphism in Japanese patients with multiple sclerosis. J Neurol Sci 2000; 179: 70–75.
21. Ponglikitmongkol M, Green S, Chambon P. Genomic organization of the human oestrogen receptor gene. EMBO J 1988; 7: 3385–3388. | PubMed | ISI | ChemPort |
22. Willing M, Sowers M, Aron D, Clark MK, Burns T, Bunten C, Crutchfield M, D'Agostino D, Jannausch M. Bone mineral density and its change in white women: estrogen and vitamin D receptor genotypes and their interaction. J Bone Miner Res 1998; 13: 695–705. | Article | PubMed | ISI | ChemPort |
23. Patel MS, Cole DEC, Smith JD, Hawker GA, Wong B, Trang H, Vieth R, Meltzer P, Rubin LA. Alleles of the estrogen receptor -gene and an estrogen receptor cotranscriptional activator gene, amplified in breast cancer-1 (AIBI1), are associated with quantitative calcaneal ultrasound. J Bone Miner Res 2000; 15: 2231–2239. | Article | PubMed | ISI | ChemPort |
24. Salmén T, Heikkinen A-M, Mahonen A, Kröger H, Komulainen M, Saarikoski S, Honkanen R, Mäenpää P. The protective effect of hormone-replacement therapy on fracture risk is modulated by estrogen receptor genotype in early postmenopausal women. J Bone Miner Res 2000; 15: 2479–2486. | PubMed |
25. Shimokata H, Ando F, Niino N. A new comprehensive study on aging — the National Institute for Longevity Sciences, Longitudinal Study of Aging (NILS-LSA). J Epidemiol 2000; 10: S1–S9. | PubMed | ChemPort |
26. Yamada Y, Ando F, Niino N, Shimokata H. Association of polymorphisms of the estrogen receptor alpha gene with bone mineral density in community-dwelling Japanese. Int J Mol Med 2002; 80: 452–460.
27. Taylor HL, Jacobs Jr DR, Shucker B, Knudsen J, Leon AS, DeBacker G. A questionnaire for the assessment of leisure-time physical activities. J Chron Dis 1978; 31: 741–755.
28. Montoye HJ. Estimation of habitual physical activity by questionnaire and interview. Am J Clin Nutr 1971; 24: 1113–1118.
29. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18: 499–502. | PubMed | ISI | ChemPort |
30. Thompson EA, Deeb S, Walker D, Motulsky AG. The detection of linkage disequilibrium between closely linked markers: RFLPs at the AI-CIII apolipoprotein genes. Am J Hum Genet 1988; 42: 113–124. | PubMed | ISI | ChemPort |
31. Rankinen T, Perusse L, Weisnagel SJ, Snyder EE, Chagnon YC, Bouchard C. The human obesity gene map: the 2001 update. Obes Res 2002; 10: 196–243. | PubMed | ISI | ChemPort |
32. Macor C, Ruggeri A, Mazzonetto P, Federspil G, Cobelli C, Vettor R. Visceral adipose tissue impairs insulin secretion and insulin sensitivity but not energy expenditure in obesity. Metabolism 1997; 46: 123–129. | Article | PubMed | ISI | ChemPort |
33. Cooke PS, Heine PA, Taylor JA, Lubahn DB. The role of estrogen and estrogen receptor-alpha in male adipose tissue. Mol Cell Endocrinol 2001; 178: 147–154. | Article | PubMed |
34. Snijder MB, Dekker JM, Visser M, Yudkin JS, Stehouwer CD, Bouter LM, Heine RJ, Nijpels G, Seidell JC. Larger thigh and hip circumferences are associated with better glucose tolerance: the horn study. Obes Res 2003; 11: 104–111. | PubMed | ISI |
35. Goessl C, Plaschke J, Pistorius S, Hahn M, Frank S, Hampl M, Gorgens H, Koch R, Saeger HD, Schackert HK. An intronic germline transition in the HNPCC gene hMSH2 is associated with sporadic colorectal cancer. Eur J Cancer 1997; 33: 1869–1874.
36. Blaszyk H, Hartmann A, Sommer SS, Kovach JS. A polymorphism but no mutations in the GADD45 gene in breast cancers. Hum Genet 1996; 97: 543–547.

Acknowledgements

We are grateful to the participants in the study. We also thank all the investigators, research assistants and laboratory technicians who have contributed to this study. This study was supported by a Grant-in-Aid for comprehensive Research on Aging and Health from the Ministry of Health, Labour and Welfare of Japan.