SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure
  8. REFERENCES

The FTO gene has significant polymorphic variation associated with obesity, but its function is unknown. We screened a population of 150 whites (103F/47M) resident in NE Scotland, United Kingdom, for variants of the FTO gene and linked these to phenotypic variation in their energy expenditure (basal metabolic rate (BMR) and maximal oxygen consumption VO2max) and energy intake. There was no significant association between the FTO genotype and BMR or VO2max. The FTO genotype was significantly associated (P = 0.024) with variation in energy intake, with average daily intake being 9.0 MJ for the wild-type TT genotype and 10.2 and 9.5 MJ for the “at risk” AT and AA genotypes, respectively. Adjusting intake for BMR did not remove the significance (P = 0.043). FTO genotype probably affects obesity via effects on food intake rather than energy expenditure.

Obesity is a chronic condition associated with increases in the risk of type 2 diabetes, hypertension, cardiovascular disease, and various forms of cancer (1). Over the past 30 years, there has been a dramatic increase in the prevalence of obesity throughout the entire world (2). In the United States, for example, levels of obesity (BMI > 30 kg/m2) in 1999–2002 were 32% of the adult population (3). Studies of mono- and di-zygotic twins have established that the variation in adult body fatness has a large genetic component (4,5,6,7).

Studies of the regulation of food intake in rodent models of obesity have led to the identification of several cases where obesity is the consequence of single loss of function mutations in particularly important genes involved in regulation of body weight (8). However, these monogenic forms of obesity appear to be very rare (9). Over the past few years, several genes have been identified from wide-scale screening studies with common variants associated with differences in obesity. These studies include variants in the INSIG (10), GAD2 (11), ENPP1 (12), and FTO (13) genes. Attempts to replicate the effects of variants in the first three of these genes have met with variable success (14,15,16). However, the effect of polymorphic variation in the FTO gene (13), which was established in a genome-wide association study of a population of 1,924 type 2 diabetics combined with 2,938 controls, was replicated in 13 separate cohorts comprising almost 39,000 individuals. This verification lends strong support to the suggestion that this gene has common variants (the AA and AT genotypes) that predispose to obesity, relative to the wild (TT) genotype. The effect of genotype on BMI was detectable from 7 years old upward (1). Subsequent studies have confirmed the effect of the FTO variants on both obesity and diabetes (17,18).

The FTO gene was originally cloned from a mutant mouse that had fused toes (Ft) (19) although the mutation in that instance included deletion of at least six separate genes, one of which was FTO. Expression studies indicate that FTO is widely expressed in many tissues, but has its highest expression in the brain, particularly the arcuate nucleus of the hypothalamus (20). A close-by gene, KIAA1005, is also highly expressed in the hypothalamus (13). Frayling et al. (13) concluded that the 47-kb intronic segment of FTO is likely to contain the predisposing variant(s); however, there was no mechanism to explain how these polymorphisms alter function or expression of FTO, KIAA1005, or more distant genes. The central distribution of FTO and KIAA1005 indicates that the polymorphic variation in FTO may play a causal role in the regulation of either energy intake and/or energy expenditure, both of which are controlled by neuropeptides in the hypothalamus (21,22). Observation that expression of FTO is altered by feeding and fasting in wild-type mice supports this view (20) although recent studies indicate a potential role in lypolysis (23).

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure
  8. REFERENCES

To explore the functional basis of the common variants of the FTO gene, we explored the association between the FTO genotype and phenotypic variation in a cohort of 150 adult whites resident in NE Scotland, United Kingdom. We have extensive phenotype data for this population concerning energy balance. These data include measurements of food intake using a 7-day weighed intake protocol, enabling separation of energy intake into component macronutrients including fat, saturated and monounsaturated and polyunsaturated fats, protein, carbohydrate, and alcohol consumption. Subjects have also been screened for body composition (by dual-energy X-ray absorptiometry and whole body displacement air plethysmography) as well as by standard anthropometry (height, weight, waist and hip ratios as well as skinfolds) and basal metabolic rates (BMRs) by hood respirometry, and maximal oxygen consumption (VO2max) extrapolated from a submaximal fitness test. In addition, this population has also been measured for blood pressure, as well as circulating insulin, leptin, thyroid hormones (T3 and T4) and fasted glucose levels. We have used this population previously to dissect the contributors to variation in BMR (24,25) and established an association between polymorphic variation in the UCP2 gene and circulating leptin levels (26). To explore the association between phenotypic measures and polymorphism in FTO, we genotyped all 150 subjects for the FTO SNP rs9939609.

Subject characteristics

One hundred and fifty unrelated adults (females n = 107, males n = 43), aged between 21 and 64 years and BMI range of 16.7–49.3 kg/m2, were recruited (Table 1) by newspaper advertisement to participate in a study investigating genetic and environmental influences on body weight. The subject group is a representative sample of a Scottish population, with 1 in 5 obese (BMI > 30) and 54% of the population collectively overweight and obese (BMI > 25). Demographically, Aberdeen is almost completely white, with only minor Asian communities. We did not make any selection of participants on the basis of race, but reflecting the local demography, all the recruited participants were whites. Participants were only included if they were not on any special medical diet; had stable weight (weight change of no >2 kg in the previous 3 months); were otherwise normal, based an extensive medical examination, screening blood tests (full blood count, renal, liver, and thyroid function) and electrocardiogram, and took no regular prescribed medication, vitamin or mineral supplements. The 150 subjects were selected from 167 individuals invited to take part in the study. Of the 17 subjects not enrolled, 12 failed the selection criteria and 5 did not attend for the initial screen. The study was approved by the Joint Ethical Committee of Grampian Health Board and The University of Aberdeen. Written informed consent was obtained.

Table 1.  Phenotypic aspects related to body composition, circulating factors and energy expenditure of 150 whites living in NE Scotland divided by the common variants in the FTO gene
inline image

Phenotyping methods

We measured BMR, body composition by dual-energy X-ray absorptiometry and by skinfolds, height, weight, hip and waist circumferences, circulating leptin, insulin, T3 and T4 levels (24,25).

Maximum oxygen consumption (VO2max)

Maximal oxygen consumption (VO2 max) was estimated by the submaximal fitness test. The test comprised sedentary routines as well as several steps with increasing workloads on a bicycle ergometer (Tunturi E850, Tunturi, Finland), similar to the calibration protocol devised by Ceesay et al. (27). The workloads as follows:

    •5-min sitting

    •5-min standing

    •5-min cycling at low resistance (50 watts)

    •5-min cycling to raise heart rate (HR) (e.g., 75 watts)

    •5-min cycling to raise HR (e.g., 100 watts)

    •5-min cycling to raise HR (e.g., 125 watts)

Pedal speed remained constant at 60 r.p.m. throughout the test and resistance was gradually increased to elevate HR and VO2. During the last 2 min of each workload, when HR had adjusted and was stable, breath-by-breath VO2 and VCO2 were measured by indirect calorimetry (Vmax29 metabolic cart; SensorMedics, Yorba Linda, CA), using a mouthpiece and noseclip.

The data derived from the procedure was used to predict the subject's physical performance or maximum oxygen uptake (VO2max), by extrapolation of the regression of VO2 against HR to the subject's maximum HR. This was calculated as 220 minus age (years) (28). The VO2 that coincides with maximal HR was assumed to be the maximum oxygen uptake (VO2max) and was expressed in ml per minute per kg body weight.

Food intake

For measurement of food and macronutrient intake, subjects were asked to record all foods and drinks consumed for a consecutive 7-day period. They were provided with digital, electronic scales (Soehnle model 820; Soehnle, Melville, NY) that have a tare facility. Scales weighed to 1 kg with a resolution of 1 g. Scales were calibrated annually and checked before each measurement period at four points over the scale's range using reference weights (Thomson Scientific, Blackburn, Aberdeen, UK). Participants were also provided with a food diary for recording a description of the food or drink, time of consumption, weight of food and leftovers. They were encouraged to record all recipe formulations and to keep all packaging for ready-to-eat food products, as described by Bingham et al. (29).

Calculation of average energy and nutrient intakes by day and by meal was by a computerized version of standard food tables. Diets were analyzed using Diet 5 (The Robert Gordon University, Aberdeen), a computerized version of MCCance and Widdowson. The composition of foods and supplements (1991). Nutritional information from manufacturers was added to the Diet5 database for processed foods. Thus, total food energy and nutrient intake could be quantified. The average of the 7-day collection was used for this analysis. We screened the data on energy intake for misreporting using standard techniques (30,31).

Genotyping methods

Genomic DNA was extracted from 2 ml of whole blood using QIAamp DNA Blood Midi Kit (QIAGEN, Hilden, Germany). Amplification of a 882-bp region surrounding rs9939609 in the FTO gene was performed using the oligonucleotide primers P1:5′-TGACGGCTGTAGAGGATAGACCAT-3′ and P2: 5′-CACCGTGTTAGCCAGGATAGTTTC-3′. Cleaned PCR products (MultiScreen HTS PCR; Millipore, MA) were genotyped using primer P3: 5′-AACAGAGACTATCCAAGTGCATCAC-3′ and GenomeLab SNP-Primer Extension Kit (Beckman Coulter, CA) running products on the CEQ 8000 GeXP (Beckman Coulter, CA).

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure
  8. REFERENCES

Mean subject characteristics in our population divided by genotype are presented in Table 1 and for energy intake parameters Table 2. The population consisted of 28 individuals with the AA genotype, 70 AT, and 52 TT. Overall frequency of the A allele was 0.42. The observed distribution of genotypes did not differ significantly (χ2 = 0.147, d.f. = 2, P = 0.929) from the Hardy-Weinberg expectations which were AA (P = 0.1764) n = 26, AT (P = 0.4872) n = 70, TT (P = 0.3364) n = 51. Subject ages in our population varied between 21 and 64 years. Because age and sex covary with many features of energy balance, it was important that the genotype distribution for FTO was not significantly biased with respect to age (F = 0.01, P = 0.977: Table 1) and that there were no bias in genotype frequency with respect to sex (χ2 = 0.568, d.f. = 2, P = 0.753). In this cohort, there was no significant effect of FTO genotype on body weight (P = 0.571), height (P = 0.491), BMI (P = 0.796), waist circumference (P = 0.955), hip circumference (P = 0.487) and total body fat by dual-energy X-ray absorptiometry (P = 0.462) (Table 1). When analysis was restricted to individuals with VO2max <27 mls O2/kg/min, the effect of genotype on BMI remained nonsignificant, but the trend was reversed and in line with the original description of the gene effects on obesity susceptability (AA = 29.65, AT = 29.65, TT = 28.46). No association was detected between the FTO genotype and levels of circulating leptin or insulin (both with or without the effects of total fat mass on these hormones accounted for), T3 (P = 0.122) or T4 (P = 0.722). FTO genotype was also not related to fasting glucose (P = 0.489) or to systolic (P = 0.193) or diastolic (P = 0.194) blood pressure. In addition to the absence of any effects detected by ANOVA, no significant effects were detected when linear regression analysis was used to explore gene dosage effects of the A allele (Table 1).

Table 2.  Phenotypic aspects related to food intake of 150 whites living in NE Scotland divided by the common variants in the FTO gene
inline image

Consistent with the absence of an effect of genotype on circulating thyroid hormones, there was no effect of the FTO genotype on BMR (P = 0.544) or on VO2max (P = 0.243) (Table 1). We have previously shown that BMR in this cohort is related to lean tissue mass, fat tissue mass, and age (27). The absence of an effect of genotype remained when the variation due to these covariates was controlled (general linear model: P = 0.498). These variables also influenced VO2max, and similarly, when they were controlled, the remaining variation in VO2max was not related to genotype (P = 0.601). We therefore found no evidence that the AA and AT genotypes, which were previously linked to elevated levels of obesity relative to TT, were associated with lowered basal metabolism or maximal oxygen consumption.

We screened the data on energy intake for misreporting using standard techniques (31,32,33) and eliminated any data where the daily intake was less than a cutoff 1.3 × BMR (n = 97 (65%) subjects retained). We also analyzed the data with the cutoff at 1.2 × BMR, 1.15 × BMR and 1.10 × BMR and the following results were not sensitive to the level at which the cutoff was applied. All the above results concerning energy expenditure and other parameters were replicated in this reduced data set. In contrast to the lack of association between the FTO genotypes and energy expenditure, we found a significant effect of the genotype on daily energy intake (ANOVA: F = 3.09, P = 0.024), with the AT and AA genotypes having higher intake (10.2 MJ/day and 9.5 MJ/day respectively) than the TT genotype (9.0 MJ/day) (Table 2). Daily energy intake was significantly related to lean tissue mass (but not fat tissue mass or age) derived from dual-energy X-ray absorptiometry (linear regression: r2 = 0.366, F = 56.62, P <0.001). The effect of the FTO genotype remained significant (general linear model: F = 3.21, P = 0.045) when the effect of lean tissue mass on intake was controlled. We also adjusted energy intake for BMR, and found the effect was also still significant (F = 3.27, P = 0.043). There were no significant differences across genotypes in the intake of saturated fat (P = 0.535), monounsaturated fat (P = 0.313), polyunsaturated fat (P = 0.080) and carbohydrate (P = 0.095) (Table 2). Protein intake was significantly higher in the individuals with the AT phenotype (P = 0.020) and this effect remained significant when intake was adjusted for the effect of lean tissue mass. Linear regression analysis revealed that there was no evidence of a linear gene dosage effect for the A allele on energy intake suggesting a simple dominance effect (Table 1). A linear gene dosage effect for the A allele however did approach significance for alcohol intake.

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure
  8. REFERENCES

The lack of an effect of the FTO genotype on indices of obesity in this study was consistent with Frayling et al. (13), where no significant association between genotype and BMI was evident in samples of <1,000 subjects. Although we found no effect of the FTO gene variants on levels of obesity in our population, this was not surprising given the subtle nature of the effects reported previously and the sample size of subjects involved in the present study. Moreover, it has been reported (oral communication, American Diabetes association meeting, 23 June 2007, A.R. Schuldiner, personal communication) that the effect of FTO on BMI is not detected in subjects at high levels of physical activity or fitness (VO2max >27 ml O2/kg/min). Since our population had relatively high levels of VO2max, this may also explain why no effect on obesity levels was detected in our sample (Table 1). Indeed, in the population with fitness levels <27 mls O2/kg/min, the trend linking genotype and obesity was reversed and in line with the description in the original paper (13). This is consistent with the observation that low physical activity accentuates the genetic effect (32). The absence of a significant effect on obesity in our sample does not invalidate the search for impacts of FTO gene polymorphisms on aspects of energy balance, which may be more directly influenced by the FTO gene. The fact that we observed effects of the FTO polymorphism on food intake in such a relatively small sample suggests that the FTO gene plays a role in regulation of food intake. Because many other factors affect energy balance (such as levels of resting metabolism and physical activity), and energy balance is only one of several factors that influence BMI, the need for substantially larger cohorts of individuals to discern effects of FTO on BMI is unsurprising.

The absence of an effect of polymorphisms at SNP rs9939609 on BMR and circulating leptin levels does not accord with the recent observations that polymorphic variation at SNPs rs17817449 and rs1421085 in the FTO gene are linked to differences in circulating leptin levels and resting metabolic rate (33)—although these effects also became nonsignificant when adjusted for the significant effect of BMI. The effects of the A allele on energy intake appeared to be a simple dominance effect, as there was no evidence for a gene dosage effect via linear regression. This contrasts the incomplete dominance effect previously reported on BMI. The effect was significant if we used total energy intake or if we adjusted intake for fat-free mass, or if we adjusted for BMR. The almost significant gene dosage effect for alcohol intake was interesting, but it seems unlikely that this can be an important mode of action for FTO because the effect on BMI is significant in cohorts of children (13,17) where alcohol intake as a component of daily energy demands is presumably negligible. Recent studies have shown that the FTO gene shares sequence motifs with Fe(II)- and 2-oxoglutarate-dependent oxygenases which are involved in DNA methylation (20). Gene expression studies of wild-type mice indicate that FTO is abundant in the arcuate nucleus of the hypothalamus and is regulated by feeding and fasting (20). This strongly implicates FTO as having a role in regulation of food intake, which is supported by our observations. It is important to note that in our data, we made a total of 25 separate analyses, technically therefore to reach statistical significance applying the Bonferoni correction; none of the results we reported reached this enhanced significance criterion. These results therefore remain preliminary until they can be confirmed in a larger population. The strong a priori expectation of an effect on energy intake (20), however, leads us to believe that they are probably correct, and had we only looked for this single effect in our data, the statistics would have supported that interpretation.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure
  8. REFERENCES

This study was funded by the Scottish Executive Environment and Rural Affairs Department (SEERAD). The funding body played no part in the design of the work or in the preparation of this paper. This submission was reviewed and approved by the senior management of the Rowett Research Institute where the work was performed. We are grateful to Sandra Murison, Sharon Wood, and Jackie Duncan for technical assistance with the phenotyping and genotyping. Colin Selman and Ela Krol made helpful comments on earlier drafts of the manuscript.

Disclosure

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure
  8. REFERENCES

The authors declared no conflict of interest.

REFERENCES

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. Disclosure
  8. REFERENCES