A Val1483Ile polymorphism in the human fatty acid sythase gene (FAS) has recently been shown to be associated with lower percentage of body fat and substrate oxidation rates in Pima Indians, but its role in other populations has not been described. Here, we investigate the effect of this variant on obesity in Caucasian children and adolescents.
Subjects and methods:
In total, 738 Caucasian children and adolescents aged 6–17 years of the Leipzig Schoolchildren cohort, which constitutes an unselected representative German population and 205 obese children (body mass index (BMI) 2.71±0.04 SDS) were genotyped for genotype–phenotype associations.
The frequency of the Ile-allele was lower in German Caucasians compared with Pima Indians (0.03 compared to 0.10). Using generalized linear regression models, there was no effect of the polymorphism on BMI in the whole normal population. However, we identified a significant interaction effect between sex and genotype (P=0.004). Subsequent sex stratified analyses revealed a lower BMI SDS in boys with Ile/Val genotype compared to Val/Val (−0.36±0.29 vs 0.09±0.05, P<0.05), while an opposite effect was observed in girls (0.48±0.19 vs 0.09±0.05, P<0.05). In genotype–phenotype associations in obese children, the polymorphism did not affect parameters of insulin, glucose, or lipid metabolism in the whole population. Again, however, obese boys with Ile/Val genotype had significantly higher high-density lipoprotein (HDL) cholesterol levels (1.46±0.07 vs 1.23±0.03 mmol/l, P<0.05).
In conclusion, our findings suggest a sex-specific protective effect of the Val1483Ile polymorphism in FAS for obesity in Caucasian boys. In addition, the polymorphism may be associated with a beneficial lipid profile in obese boys.
As in adults, obesity presents a major health burden in children with up to 30% of children and adolescents being overweight in industrialized countries.1, 2 Along with environmental factors, recent data suggest a strong genetic component in the susceptibility to obesity.3 Thus, the search for genes that increase susceptibility for common forms of obesity has become increasingly important.
The human fatty acid synthase gene (FAS) codes for an enzyme essential for de novo fatty acid synthesis and cellular substrate energy metabolism.4 It is positioned on chromosome 17q25 within a region of linkage for body fat content in Pima Indians.5 Inhibition of FAS induces a rapid decline in fat stores in mice, suggesting a role for FAS in energy homeostasis.6 Hence, the FAS may constitute a putative candidate gene for development of obesity. Indeed, we recently identified a novel polymorphism predicting Val1483Ile substitution in the FAS, which was a significantly associated with body fat content and substrate oxidation rate in Pima Indians.7 However, since Pima Indians represent a genetically isolated population with one of the highest prevalence rates of obesity and T2DM worldwide,8 it is of interest to investigate this FAS polymorphism in other populations to confirm its clinical relevance.
Children constitute a valuable study population since phenotype is influenced to a lesser extent by comorbidities, by comparison with adults, and gene polymorphisms may be expected to cause clinically relevant phenotypes at a young age. The aim of the present study was, therefore, to evaluate the effects of the FAS Val1483Ile polymorphism on obesity in a large representative cohort of Caucasian children and adolescents. In addition, using a mixed model analysis, we assessed the relationship between Val1483Ile and obesity in Pima Indians. Finally, we investigated the effect of this polymorphism on lipid and glucose metabolism in a cohort of obese children and adolescents with detailed metabolic testing.
Leipzig schoolchildren cohort
The cohort was derived from the Leipzig Schoolchildren Project that investigated anthropometric and clinical parameters in 2500 children aged 6–17 years from 1999 to 2001 and constitutes a representative normal population of Caucasian children.9 DNA was available in 738 children (358 boys, 380 girls).
A total of 205 Caucasian children and adolescents were consecutively recruited from the obesity clinic of the University Hospital for Children & Adolescents Leipzig. All obese children had a detailed metabolic work-up including an oral glucose tolerance test and lipid profile.
Demographic and anthropometric parameters of both cohorts are given in Table 1. In all subjects and patients, a careful history and physical exam including anthropometric measurements were obtained. Height and weight were determined using precision stadiometers and scales to the nearest of 0.1 cm and 0.1 kg, respectively. Waist-to-hip-ratio (WHR) was calculated from measurements of waist circumference at the smallest circumference between hip and chest, and hip circumference using a nonstretchable metric band. For standardization of height, weight, and body mass index (BMI) reference percentiles for central Germany (used for consensus guidelines of the German Working Group for Pediatric Obesity) were applied.10 Data for these parameters are given as absolute values and/or SD scores (SDS). According to current consensus guidelines of the German Working Group on Pediatric Obesity, children with a BMI 1.88 SDS were considered obese.
All patients and/or their guardians gave written informed assent/consent to be studied. The study was approved by the ethical committee of the University of Leipzig.
Pima Indian cohort
The cohort comprised 918 (512 diabetic, 406 nondiabetic) men (N=413) and women (N=505) aged>18 years selected from the ongoing longitudinal study of the etiology of T2DM among the Gila River Indian Community in Arizona.8 All participants provided written, informed consent, and all studies were approved by the Tribal Council and the Institutional Review Board of the National Institutes of Diabetes and Digestive and Kidney Diseases.
Genotyping of the Val1483Ile
Genotyping of the Val1483Ile was performed using the Taqman SNP Genotyping Assay (Applied Biosystems Inc.). Oligonucleotide sequences are available upon request. The TaqMan genotyping reaction was amplified on a GeneAmp PCR system 9700 (50°C for 2 min; 95°C for 10 min, and 95°C for 15 s, and 62°C for 1 min, for 38 cycles) and fluorescence was detected on an ABI PRISM-7000 sequence detector (Applied Biosystems Inc.).
Standard descriptive and comparative statistics were applied to characterize and compare clinical parameters in both sexes (Table 1). Variables that did not comply with normal distribution were log-transformed before analyses. The effects of the genetic variant on obesity-related phenotypes were evaluated using generalized linear regression models. Data were analyzed using the SPSS software package (SPSS v.11.5 Inc.; Chicago, IL, USA).
Mixed model analysis in Pima Indians
To account for the fact that many individuals attended >1 examination, the mixed models were fit with an unstructured covariance matrix that allowed for correlation in the BMI between different examinations in the same individual. Thus, in addition to the fixed effects, the mixed model estimated random effects representing the relation between different exams in the same individual. The parameters were estimated by a maximum-likelihood procedure (PROC-MIXED). Tests for genotypic association were undertaken assuming an additive effect of the alleles on the phenotype. This model is the most robust of the possible models to misspecification of the true model.11 All analyses were performed according to the procedures of the Statistical Analysis System software (SAS-Institute).11
We investigated the effect of Val1483Ile on obesity in a representative population of 738 German children and adolescents. Main anthropometric characteristics are given in Table 1. The frequency of the Ile-allele was lower in German Caucasians (0.03) as compared to Pima Indians (0.10). The genotype distribution was in Hardy–Weinberg equilibrium (P=0.38).
In whole study population of German schoolchildren, we did not detect a significant association between the Val1483Ile and BMI (P=0.64), BMI (SDS) (P=0.95), height (SDS) (P=0.42), waist (P=0.89), or WHR (P=0.43) (Table 2). Further analyses revealed, however, that the effect of the Val1483Ile on BMI-SDS was modified by an interaction of genotype with sex (P for sex*genotype interaction=0.004). Based on this observation, the analysis was stratified by sex, which allowed the identification of a significantly lower BMI-SDS in boys with Ile/Val compared to Val/Val homozygotes, while an opposite effect was observed in girls (Figure 1). No association in either sex was found for other anthropometric parameters.
We also tested the association between the Val1483Ile and BMI in Pima Indians by performing mixed-model analyses on ∼5000 observations from the ongoing longitudinal study of the etiology of obesity and T2DM among the Gila River Indian Community in central Arizona. The analysis only showed borderline association of the variant with BMI (34.8±0.3 for Val/Val vs 33.8±0.6 for Ile/x; P=0.09 in an additive mode of inheritance). However, no sex*genotype interaction was observed in this cohort (P for sex*genotype interaction=0.76).
Finally, we performed genotype–phenotype analyses in a cohort of 205 obese children and adolescents (Table 1) in order to assess the role of the polymorphism in the etiology of other metabolic phenotypes. Performing analyses in the entire cohort of obese children did not reveal any genotype-dependent differences in parameters of glucose, insulin, or lipid metabolism between the genotypes (Table 3, Figure 1). However, as in the earlier analyses, stratification by sex allowed the identification of significant associations between the Val1483Ile and high-density lipoprotein-cholesterol (Figure 1) and the low-density lipoprotein :HDL ratio in boys (2.72±0.41 for Val/Ile vs 3.57±0.10 for Val/Val, P<0.05).
We also compared frequencies of the Val/Ile genotype in this cohort of obese children and in a lean control group selected from the population of 738 schoolchildren. The lean control group comprised 508 subjects (243 boys and 265 girls; mean age of 12.0±0.12 years; mean BMI-SDS of −0.052±0.02) with BMI ranging from −1.0 SDS to 1.0 SDS. There were no statistically significant differences in genotype distribution between obese and lean control groups (OR=1.53, 95%CI (0.79–2.97), P=0.21 after adjusting for age and sex in a logistic regression analysis).
FAS constitutes a promising candidate gene for human obesity, since animal models have shown that FAS is involved in the development of obesity through the regulation of feeding behavior. In a previous study of the FAS in Pima Indians, we identified the Val1483Ile polymorphism, and reported significant associations with percent body fat and substrate oxidation rates.7 To assess the relevance of this finding in other populations, we aimed to evaluate whether this polymorphism affects obesity and related metabolic phenotypes in a representative cohort of Caucasian children. Children constitute a valuable study population since many clinically relevant phenotypes may be less influenced by existing comorbidities. Nonetheless, there was no association between the Val1483Ile with BMI, waist, or WHR in the initial analyses. However, we identified a significant sex–genotype interaction and subsequent sex stratified analyses revealed that boys with Ile/Val had significantly lower BMI than boys with Val/Val genotypes. This is consistent with previous findings in Pima Indians, in whom carriers of the Ile variant had lower percent body fat and higher lipid oxidation rates.7 In contrast to what was observed in the boys, BMI was higher in girls carrying the Ile/Val genotype as compared to girls with Val/Val genotype. Hence, it appears that the Val1483Ile variant in the FAS functions in a sex-specific manner in Caucasian children and adolescents. Sex-specificity is not uncommon in the context of obesity. Sex specific differences in association studies on polymorphisms in genes controlling obesity, glucose, and lipid metabolism have been previously reported. These include genes such as leptin12 and its receptor,13 resistin,14 and UCP3.15 It is possible that in girls, potential obesity protective effects of the Val1483 variant are masked or actually reversed via hormonally controlled mechanisms. For example, hormonally controlled effects of the PYY and Y2R genes with obesity have been reported in men, but not women.16 It is also possible that other female-specific obesity genes modify the effects of FAS polymorphism. For example, the –512C/T variant in the human winged helix/forkhead transcription factor gene (FOXC2) was shown to be associated with basal glucose turnover and fasting plasma triglycerides in women only.17, 18 Although several studies replicated sex-specific effects in FOXC2, the physiologic basis of these sex differences remains unclear. Hence, although we cannot exclude the possibility that these associations represent false-positive associations, population- or age-specific phenomenons, we propose that FAS may have a modest sex-specific effect on obesity in populations that are not morbidly obese.
The Val1483Ile variant is positioned within the interdomain region of the FAS that binds the two dynamic active centers of the FAS dimer.19, 20 Therefore, alteration of this site could potentially reduce FAS activity by changing the configuration of catalytically active FAS.
Considering that FAS is an enzyme crucially involved in cellular fatty acid synthesis and energy metabolism, such functional alterations may affect glucose and/or lipid metabolism reflected by altered lipid profiles. This hypothesis is supported by studies showing that the FAS inhibitor C75 acts both centrally to reduce food intake and peripherally to increase fatty acid oxidation, leading to rapid and profound weight loss, loss of adipose mass, and resolution of fatty liver in diet-induced obese mice.21 Furthermore, studies on FASKOL (FAS-knockout in liver) mice showed that products of FAS reaction regulate glucose, lipid and cholesterol metabolism by serving as endogenous activators of distinct physiological pools of PPARα.22 These hypotheses from animal studies were supported by recent findings in Pima Indians suggesting that lower percentage of body fat seen in subjects with the Val/Ile genotype resulted from their higher 24-h respiratory quotient, that is, they preferentially oxidize lipids over carbohydrates, which could result in a decreased body lipid mass.7 Our present data may further support this notion by showing a more beneficial lipid profile with a tendency toward lower plasma triglyceride levels and significantly higher HDL levels in obese boys, even after adjustment for BMI. Again, this effect was sex specific as no significant effects were observed in obese girls. Interestingly, we also observed a tendency toward increased liver enzymes in patients with the Val/Ile genotype, which is again in accordance with findings from animal studies of the liver-specific FAS-knockout mouse model that presented increased liver enzymes and decreased total cholesterol levels, hence resembling a phenotype similar to that seen in our children.
The lack of statistically significant associations for some of the relationships we report here could be due to a lack of statistical power, owing to the relatively low frequency of the minor Ile allele. It may, therefore, be worthwhile pursuing these genotype–phenotype associations in larger cohorts, with the aim of confirming a functional effect of the FAS on body fat content and lipid profile in humans.
In summary, our findings suggest a protective effect of the Val1483Ile mutation in FAS against the development of obesity in Caucasian boys, which appears to be sex specific. This finding is consistent with data from Pima Indians, although in the Native American family-based study there was no association to sex. In addition, polymorphisms in the FAS may influence lipid metabolism in obese boys.
We thank all those who participated in the studies. We appreciate the help of the nurses and physicians who performed the clinical examinations and data collection. The help of Roy Tauscher for DNA extraction and consenting is highly appreciated. This work was supported by grants from the Deutsche Forschungsgemeinschaft (KFO-152 to AK and MS) from the Interdisciplinary Center of Clinical Research at the University of Leipzig to AK (B21) and to MS (Z14), from the European Community ‘PIONEER’ (to WK), and from the German Diabetes Association to AK and to PK. The Leipzig schoolchildren project was supported by unrestricted grants from Pfizer Pharma GmbH and Novo Nordisk GmbH to WK. The studies in Pima Indians were supported via the NIDDK Intramural Research Program.
About this article
Nature Reviews Endocrinology (2012)