| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Nutrition Sciences (C.L.-C., B.A.G., J.R.F.) and Human Studies (G.R.H.), and Clinical Nutrition Research Unit (G.R.H., B.A.G., J.R.F.), University of Alabama, Birmingham, Alabama 25294; and Pennington Biomedical Research Center (J.C.L.), Baton Rouge, Louisiana 70808
Address all correspondence and requests for reprints to: Dr. Cristina Lara-Castro, Department of Nutrition Sciences, 1675 Webb Nutrition Sciences Building, Room 244, University of Alabama, Birmingham, Alabama 35294. E-mail: larac{at}uab.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Even though the association of the Ala54Thr polymorphism and body fat phenotypes has been previously explored, less is known about its association to body fat distribution phenotypes, particularly abdominal fat accumulation. Body fat distribution, especially central abdominal fat accumulation, is associated with increased risk of metabolic and cardiovascular disease. The vast majority of the literature suggests that VAT is the single stronger marker of metabolic risk (10, 11); however, there is evidence supporting the potential contribution of abdominal sc adipose tissue (SAAT) as a predictor of risk (12).
Studies of the association of the Ala54Thr polymorphism and the abdominal fat phenotype are scarce; negative results have been reported for the association of this polymorphism and waist circumference in Finnish subjects (13, 14), and only one study in Japanese men has reported on the association of the polymorphism and VAT (15).
Functional studies of the Ala54Thr polymorphism indicate that the threonine-containing intestinal FABBP has a greater in vitro affinity for long-chain fatty acids than the alanine-containing protein, suggesting that the substitution may potentially result in increased absorption of fatty acids, with subsequent alterations in glucose metabolism, lipid processing, and body composition (16).
In light of the potential physiological role of the Ala54Thr FABP2 polymorphism and its possible ethnic-specific role in different disease phenotypes, we were interested in determining the relationship between abdominal fat compartments and the Ala54Thr FABP2 polymorphism in a biethnic group of premenopausal women. We hypothesized that there is an association between VAT and SAAT and the Ala54Thr FABP2 polymorphism in premenopausal women, and that the strength of the association between the polymorphism and the abdominal fat phenotypes in women differs by ethnic group.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A total of 223 women participated in the study. Subjects belonged to three ongoing longitudinal studies conducted in African-American and Caucasian premenopausal women: two studies on the Role of Metabolism in the Etiology of Obesity (ROMEO and JULIET; n = 100) and the Healthy Transitions Study (HTS; n = 123), which focuses on body composition changes during the transition to menopause. Only baseline data (before the intervention) were used for the present analysis. No previous or undergoing intervention that affected body composition of the subjects occurred at the time of this first evaluation. Subjects in the ROMEO study included two groups of healthy African-American and Caucasian women: one obese and one never obese. The never-obese group was comprised of women with no personal or familial history of obesity. Ascertainment for the obese group required a familial history of obesity and a BMI between 27–30 kg/m2. Subjects in the JULIET study had a BMI between 27–30 kg/m2 and a familial history of obesity, but were otherwise healthy Caucasian and African-American women. All subjects in ROMEO and JULIET were recruited by university-wide newspaper advertisement in the Birmingham, AL, area, were nonsmokers, were not taking any medication known to alter body composition (including hormones), were sedentary (defined as exercising less than once per week for the past year), were of overall good health, and had regular menstrual cycles. Similarly, HTS subjects were recruited by advertisement and word of mouth from the Baton Rouge, LA, area. To be eligible for the study, the women had to be healthy, had to be 43 yr of age or older, and had to have had five menstrual periods or more in the 6 mo before screening. Women were excluded if they were taking medication regularly (including hormones), were not having regular menstrual cycles, or had clinically abnormal results from laboratory tests or physical examination. Women in this study were recruited regardless of BMI. In all studies race was determined by self-reported African-American or Caucasian ancestry in both parents and grandparents. The institutional review board for human subjects-approved informed consent was obtained before participation in the study in compliance with the Department of Health and Human Services Regulations for Protection of Human Research Subjects.
Study design
Eligible subjects participating in these studies were admitted to the research units of University of Alabama and Pennington Biomedical Research Center and underwent indicated measurements. Specific methodology is described below.
Study procedures
Genotyping. The Ala54Thr polymorphism in the FABP2 promoter was analyzed by PCR amplification of a 170-bp fragment of the FABP2 gene encompassing the site of the polymorphism using upstream primer 5'-ACA-GGT-GTT-AAT-ATA-GTG-AAA-AG-3' and downstream primer 5'-TAC-CCT-GAG-TTC-AGT-TCC-GTC-3'. After amplification, digestion with HhaI (20.000 U/ml; New England Biolabs, Mississauga, Canada) was performed. The PCR product was incubated with the enzyme at 37 C for 3 h and then inactivated at 65 C for 20 min. Cleavage of the amplified fragment produced two fragments of 96 and 74 bp, and the Ala54Thr substitution abolished the HhaI restriction site. Expected sizes were 96 and 74 bp for normal homozygous; 170 bp for Thr homozygous; and 170, 96, and 74 bp for heterozygous.
Instrumentation. Fragment amplification was carried out by PCR technique using the Peltier Thermal Cycler (model PTC-200, MJ Research, Waterford, MA). Electrophoresis and sizing of the resulting DNA fragments were conducted using the ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Foster City, CA) and dye-labeled reverse primer (Invitrogen Life Technologies, Carlsbad, CA). For these analyses, 1 µl of the digested sample was added to a denaturing mixture containing 0.5 µl 1:5 diluted Genscan-500 TAMRA size standard (Applied Biosystems) and 12 µl deionized formamide. The mixture was heated at 95 C for 5 min and immediately chilled for 2 min. The ABI PRISM capillary electrophoresis conditions included 5-sec injections and electrophoresis for 20 min at 60 C using the GS POP-4 denatured C module setting.
Body composition analysis
For ROMEO and JULIET subjects, fat and lean mass were determined by dual energy x-ray absorptiometry (DXA; DPX-L, Lunar Radiation Corp., Madison, WI) with the use of software version 1.33 (Lunar Radiation Corp.). Cross-sectional areas of VAT and SAAT were determined by computed tomography (CT) with the use of a HiLight/HTD Advantage scanner (General Electric Co., Milwaukee, WI) set at 120 kVp (peak kilovoltage) and 40 mA. Subjects were examined in the supine position with their arms stretched above their heads, by taking a 5-mm scan for 2 sec at the level of the fourth and fifth lumbar vertebrae. The attenuation range for adipose tissue was –30 to –190 Hounsfield units according to the procedures established by Kvist et al. (17). Cross-sections of adipose tissue were determined using a computerized fat tissue-highlighting technique. VAT and SAAT were measured by separating adipose tissue areas by encircling the muscle wall surrounding the abdominal cavity with a cursor. Tissue cross-sections between –30 and –190 Hounsfield units in the respective areas were considered to be VAT and SAAT. Both intra- and interobserver test-retest reliability had an r value of 0.99 with a coefficient of variation of less than 2% on the basis of reevaluation of 20 scans.
For the subjects in the HTS, fat and lean masses were determined by DXA (Hologic QDR2000, Hologic, Inc., Waltham, MA). In addition, all subjects had an abdominal CT scan at the level of the interspace between the fourth and fifth lumbar vertebrae (10 mm thick) for determination of abdominal fat distribution (GE High Speed Advantage, GE Medical Systems). Images were stored on digital tape for analysis at the Pennington Center using the Analyze software package (CNSoftware, Rochester, MN) run on a Sun Sparc 20 workstation (Sun Microsoft, San Jose, CA). Total abdominal adipose tissue (TAT) was defined as the sum of adipose tissue pixels inside a line tracing of the skin. VAT was segmented by drawing a line around the interior of the peritoneal cavity and summing all adipose tissue pixels within this area. The difference between TAT and VAT was considered to represent SAAT. A single reader performed all image analysis. The coefficients of variation for measures of VAT and SAAT in the laboratory were 10.5% and 9.8%, respectively (repeat measures on the same subjects made 1–2 wk apart). For determination of VAT and SAAT in all study subjects, a range between –30 and –190 Hounsfield units for the upper and lower boundaries was used.
Statistical analysis
Differences in outcomes of interest between racial groups were compared by t test. Data analysis was conducted using SAS statistical software (version 8.02, SAS Institute, Inc., Cary, NC). Phenotypic variables for the present analysis were total body fat, TAT, VAT, and SAAT. Because regression models were used for all hypotheses testing, variable transformation was based on the normal distribution of residuals for models predicting each phenotypic variable. Consequently, log transformation was required for VAT and SAAT for all statistical analyses; thus, all results including these variables were derived from log-transformed data. For the VAT and SAAT models, the covariate used was log total fat mass. References to VAT and SAAT as dependent variables hereafter indicate the transformations described. No transformation was necessary for the dependent variables total body fat and TAT. The 
2 statistic was used to determine allele differences between racial groups. Regression analyses were conducted using the phenotype variables total body fat, TAT, VAT, and SAAT as dependent variables. These variables were regressed using genotype or presence of the mutant allele (marker) as predictors. Regression models based on genotype and presence of the marker will hereafter be referred to as genotypic and allelic analyses, respectively. For the genotypic analysis, every genotype was assigned a dummy code, where (0–0) represented normal homozygous, (1, 0) represented heterozygous, and (0–1) represented mutant homozygous. The intercept was set as the homozygous normal group and compared against the other two genotypes. Because the number of homozygous carriers was low (14 women, four African-American and 10 Caucasian), which limited the comparisons of this group and the other two genotypes, subjects homozygous and heterozygous for the mutation were pooled into one group, representing carriers of the marker, and compared against noncarriers of the marker. Therefore, for the allelic analysis, two groups were defined: carriers of the marker (1) and noncarriers of the marker (0). For this analysis the different phenotypes were regressed based on the presence or absence of the marker. Lean body mass and total body fat were used as covariates for the models where indicated. Indicated below are the different models tested: model 1, total fat mass = lean body mass + genotype/marker; model 2, TAT = total body fat + genotype/marker; model 3, log_VAT = log_total body fat + genotype/marker; and model 4, log_SAAT = log_total body fat + genotype/marker.
For all analyses, P < 0.05 was considered statistically significant. A post hoc analysis by study site was also conducted.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
and were compared between study sites (data not shown). Subjects in the HTS were significantly older (43 ± 8 vs. 40 ± 7 yr; P < 0.01) than subjects in the Birmingham cohort.
|
. Genotypic frequencies of the marker for both ethnic groups were in Hardy-Weinberg equilibrium (tested by 2 test, all P < 0.05). Genotype frequencies for the entire group were 60%, 37%, and 3% for the Ala/Ala, Ala/Thr, and Thr/Thr genotypes, respectively. The presence of the mutant Thr allele was 21% in the total sample; 90 of 223 subjects were carriers of the allele (83 heterozygous and seven homozygous). The frequency of the Thr allele tended to be higher in Caucasians compared with African-American women (24% vs. 19% respectively; 2 = 0.95; df = 1; n = 223; P = 0.32).
|
and 4, respectively. There were no significant differences in the physical characteristics given in Table 3 within African-American or Caucasian women carriers and noncarriers of the mutation. When the two ethnic groups were compared, African-American women carriers of the mutation had significantly higher lean body mass than their Caucasian counterparts (41 ± 6 vs. 38 ± 4 kg; P < 0.05). Caucasian women noncarriers of the mutant allele were significantly older (43 ± 8 vs. 40 ± 7 yr; P < 0.05) and had lower BMI (26 ± 4 vs. 28 ± 4; P < 0.05) than African-American women noncarriers of the mutation.
|
|
. From these analyses, subjects heterozygous for the FABP2 mutation had significantly lower levels of TAT and SAAT, after adjusting for total body fat, than normal homozygous subjects (P < 0.003 and P < 0.004, respectively); when analyses were conducted as allele presence/absence, similar results were found (Table 5). Carriers of the mutation had significantly lower TAT than noncarriers of the mutation after adjusting for total fat mass (P = 0.01). SAAT was also significantly and negatively associated with the presence of the mutant allele (P = 0.02). As indicated by the difference in the r2 values of the full models (including total fat mass and genotype) and the simple models (only total body fat), approximately 5% and 2% of the variances in TAT and SAAT respectively, were explained by the FABP2 polymorphism.
|
). Table 6 shows parameter estimates for the different fat phenotypes, by genotypic and phenotypic analyses, by race. In these models, Caucasian women heterozygous for the mutation had significantly lower TAT and SAAT than normal homozygous women (P = 0.001 and P = 0.007, respectively). Furthermore, the allelic analyses showed comparable results; that is, carriers of the mutation (mutant homozygous and heterozygous grouped together) had significantly lower levels of TAT and SAAT than noncarriers of the mutation (P = 0.004 and P = 0.03, respectively).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Only one study conducted in Japanese men has previously reported an association between the Ala54Thr FABP2 polymorphism and abdominal fat levels, specifically VAT (15). In this study, Yamada et al. (15) reported higher levels of VAT, as measured by ultrasound, in Thr homozygotes. Conversely, findings from our study did not support an association of this polymorphism and visceral adiposity in women. The reasons for these disparities may rely on differences between the populations under study. First, Japanese populations have shown marked heterogeneity in body composition compared with other populations, and second, the study by Yamada et al. included only men.
Our study shows associations of the –54 Thr allele and abdominal fat phenotypes in opposite directions to those reported previously by Yamada et al. (15). The functional role of the Ala54Thr substitution on the FABP2 is not clearly understood. As suggested by in vitro functional studies, carriers of the Thr allele are presumed to express increased intestinal absorption of fatty acids and elevated postprandial chylomicron concentrations in response to dietary long-chain fatty acids (18), although this function has not been confirmed in vivo (19). Interestingly, the –54 Thr allele in our study was not associated with increased VAT levels, as would be expected based on the in vitro studies. Gender-dependent differences in the association of the FABP2 polymorphism and fat metabolism have been recently reported (20), suggesting that this polymorphism may have a gender-dependent role in body fat distribution as well, possibly mediated through differences in the hormonal milieu of men and women. This possibility may partially explain the opposing results of the study in Japanese men and our study. We could also hypothesize that although FABP2 is only expressed at the intestinal level, there is a possibility that this polymorphism may be in linkage disequilibrium with a marker in another gene of the FABP family, which confers opposite functional effects in relation to fat homeostasis. In support of this hypothesis, Damcott et al. (2) recently reported linkage disequilibrium, and postulated opposing functional effects in lipid processing, between the Ala54Thr polymorphism and the FABP2 heterozygous haplotype in a group of nondiabetic subjects from the San Luis Valley Diabetes Group.
Abdominal fat depots are strongly influenced by genetic factors, with maximal heritability (>55%) for VAT and TAT, and greater than 40% heritability for SAAT (21). Results from this association study suggest that women carriers of the Thr allele are less likely to accumulate TAT, particularly SAAT, compared with noncarriers of the mutation. Our findings raise the question of whether sc abdominal fat in our model is a surrogate measurement of total body sc fat or fat in other regions not determined in this study. It is not likely that SAAT is surrogate of total body sc fat, because a correlation between total body fat and the polymorphism was not observed. Because approximately 70% of total body fat is sc, we may expect to find a certain degree of association between total body fat and this genetic variant. Furthermore, when we adjusted the model for peripheral fat (leg fat) instead of total body fat (data not shown), the significant association between the FABP2 polymorphism and sc abdominal fat was still present, suggesting that the association is specific for the abdominal sc depot. In contrast, it is possible that variations in the distribution of other fat depots not measured in this study (i.e. intramyocellular compartment) make it possible to find a negative association between this polymorphism and SAAT. Unfortunately, we were unable to determine associations of this polymorphism and SAAT subcompartments (superficial and deep regions), because data from these regions was available for only half the subjects. These results also justify the study of the relationship between other fat depots and the Ala54Thr FABP2 polymorphism.
The strengths of the present study are the accurate estimates of the phenotypes by standard DXA and CT imaging, the wide range of the phenotypic variables assessed, and a sample size adequate to determine associations between phenotypes and genotypes. The use of covariates, such as lean body mass and fat mass, in the different models tested allowed controlling for the confounding effects of body frame on total body fat and of fat mass in abdominal fat and abdominal fat compartments. A strength of the study is that only premenopausal women were included, which limited the confounding effect of menopause on abdominal fat accumulation. Limitations of this study included the heterogeneity of sampling, the inability to prove causation that characterize a cross-sectional study design, as well as the absence of environmental measures that could have confounded the genetic association. Furthermore, the temporal relation between the mutation and current body fat distribution should be considered. In addition, the small sample size of the study brings concerns about possible type I errors and questions the power of detection in those cases where an association was not observed. Another potential limitation of the study was the use of different DXA instruments for the assessment of total body fat and lean fat mass at the two study sites. To explore the possible influences of these two different measures in our results, we conducted post hoc separate analyses by study site and also used study site as the adjusting variable in all models tested. These analyses showed comparable regression models for the different phenotypic variables tested, for both genotypic and allelic models. These analyses indicated that the different models were reproducible in the two populations and therefore gave us justification for pooling the entire subject sample to increase the sample size and subsequently the power to detect the associations under study.
This study involved the testing of individual, rather than combinations of, genetic variants. Because variations at multiple genetic loci contribute to obesity and obesity-related traits, it is likely that allelic effects at some loci may be amplified in the presence of variants at other loci. In fact, it has been demonstrated in rodents that a combination of genetic defects increases the level of obesity significantly compared with that attributed to the single gene defect. Therefore, additional epistactic models may be useful in determining additive roles of defects at different loci in the etiology of the abdominal fat phenotypes.
We conclude that carriers of the Thr allele exhibited lower TAT and SAAT after adjustment for total body fat. This association was strong in Caucasian women and was not present in African-American women. The results of this study support a role for genetic variation of the FABP2 gene in abdominal fat phenotypes as well as the hypothesis that differences in the architecture of body fat distribution in different ethnic groups may have a genetic basis.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Stouffer’s Lean Cuisine entrees were provided by the Nestle Food Co. (Solon, OH), and Smart Ones entrees were provided by H. J. Heinz Frozen Foods (Pittsburgh, PA).
First Published Online November 30, 2004
Abbreviations: BMI, Body mass index; CT, computed tomography; DXA, dual energy x-ray absorptiometry; FABP2, fatty acid-binding protein-2; HTS, Healthy Transitions Study; ROMEO and JULIET, Role of Metabolism in the Etiology of Obesity; SAAT, sc abdominal fat; TAT, total abdominal adipose tissue; VAT, visceral abdominal fat.
Received April 7, 2004.
Accepted November 22, 2004.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
