Introduction

Genetic factors play an important role in the development of obesity [1], but the underlying genetic causes are poorly understood [2]. Two independent studies using different approaches recently demonstrated a strong association between variants in the gene fat mass and obesity associated (FTO) and obesity [3, 4]. Frayling et al. [3] found the association between FTO variants (represented by rs9939609) and obesity in a genome-wide scan for type 2 diabetes and replicated it in a total of more than 38,000 individuals. The association between SNPs in the FTO gene and type 2 diabetes was shown to be due to obesity [3]. In agreement with these findings, Dina et al. [4] reported a set of SNPs in the FTO gene (rs1421085 and rs17817449) that is consistently strongly associated with early-onset childhood and severe adult obesity. Both groups replicated the association between FTO and obesity in large cohorts [3, 4]. Moreover, recently Scuteri et al. [5] replicated the significant association between rs9930506 in the FTO gene and BMI, body weight and hip circumference in European and Hispanic Americans.

The Fto gene was originally described in a mouse with fused toes (Ft mutation) [6, 7]. Homozygous Ft mutation is embryonic lethal [6]; heterozygous mice have fused toes and hyperplasia of the thymus most likely due to impaired apoptosis [7]. Interestingly, obesity has not been reported in these mice. However, the Ft mutation consists of at least six deleted genes and no conclusions can be drawn from this model regarding the function of FTO and the relationship between the Fto gene and obesity [2]. Human de novo duplication on chromosome 16, which includes the FTO locus, has been shown to be associated with mental retardation, obesity, dysmorphic facies and severe bilateral clinodactyly of some fingers [8]. Recent bioinformatics analysis revealed that FTO shares sequence motifs with Fe2+- and 2-oxoglutarate-dependent oxygenases [9]. It has also been shown that FTO localises to the nucleus and that recombinant murine FTO catalyses the demethylation of 3-methylthymine in single-stranded DNA [9].

After the association with human obesity was discovered, the name was changed to ‘fat mass and obesity associated’ (FTO) [2], but the function of the FTO gene is still unknown. FTO gene is expressed in most tissues, including adipose tissue, beta cells and brain [3]. The highest FTO expression levels have been detected in brain [3, 9].

Functional studies are necessary to define the biological role of FTO and the major FTO-‘sensitive’ organs. We therefore tested the hypothesis that FTO mRNA expression in visceral and subcutaneous adipose tissue is linked to obesity in a cohort of individuals with wide ranges of obesity and body fat distribution. In addition to FTO, retinitis pigmentosa GTPase regulator interacting protein 1-like (RPGRIP1L, also known as KIAA1005) is also thought to be involved in the development of obesity because of its genomic position near the FTO gene [3, 5]. The transcription start site of RPGRIP1L is situated only 200 base pairs from the 5′ end of FTO and ∼61 kb from the 47 kb interval containing the BMI associations. RPGRIP1L is ubiquitously expressed with relatively high levels in hypothalamus and islet. We therefore also examined RPGRIP1L mRNA expression in visceral and subcutaneous adipose tissue and its relation to obesity.

In addition, we investigated whether previously published obesity-associated FTO variants (represented by rs8050136 in our study) might explain potential variability in FTO and RPGRIP1L mRNA expression.

Methods

Participants and measurement of participant characteristics

Paired samples of visceral and subcutaneous adipose tissue were obtained from 55 Europid men (n = 26) and women (n = 29) who underwent open abdominal surgery for elective cholecystectomy or explorative laparotomy. Using OGTTs, we identified individuals with type 2 diabetes (n = 19) or normal glucose tolerance (n = 36). All participants had stable weight with no fluctuations of >2% of body weight for at least 3 months before surgery. Patients with malignant diseases or any acute or chronic inflammatory disease as determined by a leucocyte count >7,000 × 109/l, C-reactive protein >50 mg/l or clinical signs of infection were excluded from the study. Samples of visceral and subcutaneous adipose tissue were immediately frozen in liquid nitrogen after explantation. The study was approved by the ethics committee of the University of Leipzig. All participants gave written informed consent before taking part in the study.

Assays and measures of glucose metabolism and body fat content

Basal blood samples were taken after an overnight fast. Plasma insulin was measured with an enzyme immunometric assay for an automated analyser (Immulite; Diagnostic Products, Los Angeles, CA, USA). Plasma leptin levels were assessed by radioimmunoassay (Linco Research, St Charles, MO, USA). OGTT was performed after an overnight fast with 75 g standardised glucose solution (Glucodex Solution 75 g; Merieux, Montreal, QC, Canada). Insulin sensitivity was assessed with the euglycaemic–hyperinsulinaemic clamp method [10]. Percentage body fat was measured by dual-energy X-ray absorptiometry. In addition, abdominal visceral and subcutaneous fat area was calculated as previously described [11, 12].

Analysis of human FTO and RPGRIP1L mRNA expression

Human FTO and RPGRIP1L mRNA expression was measured by quantitative real-time RT-PCR in a fluorescent temperature cycler using the TaqMan assay; fluorescence was detected on an ABI PRISM 7000 sequence detector (Applied Biosystems, Darmstadt, Germany). Total RNA was isolated from paired subcutaneous and visceral adipose tissue samples using TRIzol (Life Technologies, Grand Island, NY, USA), and 1 µg RNA was reverse transcribed with standard reagents (Life Technologies). From each RT-PCR, 2 µl was amplified in a 26 µl PCR using a kit (Brilliant SYBR Green QPCR Core Reagent Kit; Stratagene, La Jolla, CA, USA) according to the manufacturer’s instructions. Samples were incubated in the sequence detector for an initial denaturation at 95°C for 10 min, followed by 40 PCR cycles, each cycle consisting of 95°C for 15 s, 60°C for 1 min and 72°C for 1 min. The following primers were used: human FTO 5′-CTGGCCAGTGAAAGGGTCTAAT-3′ (sense) and 5′-GGCAGCAAGTTCTTCCAAAGC-3′ (antisense); human RPGRIP1L 5′-GGTATTGGCAAGCTCAGGGTAA-3′ (sense) and 5′-TCAAGCCTCCAAGTCATCTCTGT-3′ (antisense). SYBR Green I fluorescence emissions were monitored after each cycle. Human FTO and RPGRIP1L mRNA expression was calculated relative to the mRNA expression of 18S rRNA, determined by a premixed assay on demand for human 18S rRNA (Applied Biosystems, Darmstadt, Germany). Amplification of specific transcripts was confirmed by melting curve profiles (cooling the sample to 68°C and heating slowly to 95°C with measurement of fluorescence) at the end of each PCR. The specificity of the PCR was further verified by subjecting the amplification products to agarose gel electrophoresis.

Genotyping of the rs8050136 FTO variant

The SNP rs8050136 was selected for genotyping as a proxy for SNPs belonging to one linkage disequilibrium (LD) group including the previously reported rs9939609 [3]. The choice of rs8050136 was based on the better quality of genotyping compared with rs9939609. Genotyping of rs8050136 was performed using the TaqMan allelic discrimination assay (Applied Biosystems) on an ABI PRISM 7500 sequence detector (Applied Biosystems). To assess genotyping reproducibility, a random ~10% selection of the samples was re-genotyped in all SNPs with 100% concordance. Genotype data were consistent with Hardy–Weinberg equilibrium (p > 0.05).

Statistical analyses

Data are shown as means ± SD unless stated otherwise. Before statistical analysis, non-normally distributed parameters were logarithmically transformed to approximate a normal distribution. Expression differences between visceral and subcutaneous adipose tissue were assessed using the paired Student’s t test. Linear relationships were assessed by least square regression analysis. Multivariate linear relationships were assessed by a general linear model. Associations between rs8050136 and FTO mRNA expression levels were assessed using generalised linear regression models under the additive and dominant/recessive mode of inheritance. Statistical analysis was performed using SPSS version 12.0 (SPSS, Chicago, IL, USA). We considered p values <0.05 to be statistically significant.

Results

Participant characteristics

The study population comprised 55 Europid volunteers (26 men, 29 women) and was characterised by a mean age of 61 ± 15 years, mean BMI of 34.3 ± 9.5 kg/m2 (women: 37.7 ± 10.7 kg/m2; men: 30.4 ± 6.1 kg/m2, p < 0.05), mean waist circumference of 119 ± 26 cm (women: 121 ± 28 cm; men: 115 ± 23 cm) and mean body fat mass of 36 ± 11% (women: 40.6 ± 11.1%; men: 30.8 ± 8.4%, p < 0.05). Analysis of 55 paired samples of visceral and subcutaneous adipose tissue revealed threefold higher FTO mRNA levels in subcutaneous than in visceral adipose tissue, while RPGRIP1L mRNA expression was 1.6-fold higher in visceral than in subcutaneous adipose tissue (Fig. 1a,b). FTO and RPGRIP1L mRNA gene expression was not different between men and women in either fat depot. Moreover, subcutaneous and visceral FTO mRNA expression were significantly correlated (Fig. 1c).

Fig. 1
figure 1

FTO mRNA and RPGRIP1L mRNA expression in individuals with wide ranges of obesity, body fat distribution, insulin sensitivity and glucose tolerance. Men, white bars/symbols, n = 26; women, black bars/symbols, n = 29. a FTO mRNA and b RPGRIP1L mRNA expression in visceral and subcutaneous (SC) adipose tissue. Values are means ± SEM. p = 0.01. c Correlation between visceral and SC FTO mRNA expression (n = 55). p = 0.016; r 2 = 0.11. d Correlations between visceral and (e) SC FTO mRNA expression and BMI. p<0.001; r 2 = 0.27 (d); p = 0.018; r 2 = 0.11 (e). Human FTO and RPGRIP1L mRNA expression was calculated relative to the mRNA expression of 18S rRNA. Data were log transformed to achieve normal distribution. AU, arbitrary units

Full size image

Associations of FTO expression

Visceral FTO mRNA was significantly correlated with BMI (Fig. 1d), per cent body fat, sex and age, but not with visceral fat area (Table 1). Univariate regression analyses revealed further significant correlations between subcutaneous FTO mRNA expression and BMI (Fig. 1e), per cent body fat, subcutaneous fat area, age, sex and leptin (Table 1). The relationship between FTO mRNA expression and BMI was significant independently of sex. However, in our study population the most obese participants were also the youngest and upon statistical adjustment for BMI, the age relationship disappeared (Table 2). Since the univariate correlation with BMI was much stronger, the age effect was most probably secondary to BMI. In a stepwise regression analysis age did not enter the model after BMI at a p level of 0.05. There was no correlation between leptin and adiponectin serum concentrations, and visceral or subcutaneous FTO mRNA expression beyond the associations of these adipokines with BMI or per cent body fat.

Table 1 Univariate linear regression between anthropometric and biochemical parameters and visceral and subcutaneous FTO mRNA and RPGRIP1L mRNA expression
Full size table
Table 2 Multiple regression analysis between age, sex, BMI and FTO mRNA expression in visceral and subcutaneous adipose tissue
Full size table

The association between FTO expression and BMI remained significant upon adjustment for age and sex in the visceral (p = 0.003), but not in the subcutaneous depot (p = 0.08; Table 2). Since no functional information is available on how FTO might contribute to human obesity, we also modelled the data by reverse logic, i.e. BMI as dependent variable (Table 3). Interestingly, visceral FTO mRNA expression remained an independent predictor of BMI even after adjustment for subcutaneous FTO mRNA expression.

Table 3 Multivariate linear regression analysis of visceral and subcutaneous FTO mRNA expression as predictor of BMI
Full size table

There were no significant relationships between both visceral and subcutaneous FTO expression and fasting plasma glucose concentration, HbA1c and parameters of insulin sensitivity, including fasting plasma insulin concentrations and glucose uptake during the steady state of a euglycaemic–hyperinsulinaemic clamp (Table 1).

In contrast to FTO, no correlation of RPGRIP1L mRNA expression with anthropometric and metabolic parameters was seen in either fat depot (Table 1).

Adipose tissue FTO mRNA expression in function of rs8050136

The SNP rs8050136 was selected for genotyping as a proxy for SNPs belonging to one LD group based on HapMap Rel 21a/phaseII (http://www.hapmap.org/cgi-perl/gbrowse/hapmap_B36/). This LD group includes rs9939609, which is a common variant previously shown to be strongly associated with obesity [3]. No significant association of rs8050136 with visceral or subcutaneous FTO mRNA levels was found (Table 4). However, despite the lack of statistical power due to the small sample size, rs8050136 was still associated with obesity in our study group (data not shown).

Table 4 FTO and RPGRIP1L mRNA expression in visceral and subcutaneous adipose tissue grouped by rs8050136 genotypes
Full size table

Discussion

Two independent studies recently revealed an association between common variants in the FTO gene on chromosome 16q12.2, which have been confirmed in more than 20 independent cohorts [3–5], and BMI. FTO is highly expressed in brain, but also detectable in most tissues, including adipose tissue and pancreatic beta cells [3, 9]. The function of the FTO gene product and its potential role for human obesity, however, are completely unknown at the moment. It is possible that only central mechanisms related to the control of appetite and satiety are involved. On the other hand, adipose tissue represents a primary candidate for studying the functional relevance of FTO in obesity. Conceivably, expression differences between lean and obese individuals could contribute to our understanding of the FTO gene product. Based on its genomic position near FTO, as well as on similar expression profiles, RPGRIP1L is also considered a plausible candidate gene involved in the pathophysiology of obesity [3, 5].

In this study, we provide first descriptive information on the relationship between FTO and RPGRIP1L mRNA expression in adipose tissue and simple anthropometric parameters including measures of obesity.

FTO mRNA expression was threefold higher in subcutaneous than in visceral fat. This difference could be due to a number of factors usually differentiating these two depots including gene transcription, insulin sensitivity and lipolysis, altered adipokine expression and others [13, 14]. Decreasing subcutaneous fat mass does not significantly improve obesity-associated metabolic abnormalities [15], whereas reduction of visceral fat mass has been shown to have significant beneficial effects on glucose metabolism and insulin sensitivity [16]. Our data indicate that fat depot differences in FTO expression may contribute to the specific properties of visceral adipose tissue. The observation that FTO expression in the two depots was only modestly correlated suggests that in a given individual up- or downregulation in one vs the other depot is controlled locally rather than in a coordinated fashion by systemic hormonal, neuronal and/or nutritional signals.

Independently of the fat depot, we found significant negative correlations between FTO gene expression and BMI or per cent body fat. Therefore our data suggest that FTO gene expression might be downregulated in response to fat accumulation. As indicated by the models (Table 2), this inverse relationship was significantly stronger for the visceral depot. From our correlation analysis it is impossible to extricate the direction of causality. Nevertheless, since genetic variation in FTO has been shown to affect the risk of obesity, it is legitimate to test the hypothesis of whether FTO expression predicts BMI. We identified visceral FTO mRNA expression as a strong predictor of BMI independently of age and sex, and even independently of FTO mRNA expression in the subcutaneous depot. This suggests that especially viscerally expressed FTO may contribute to obesity.

Obviously, the number of participants in our study was relatively small and the association between BMI and adipose tissue FTO mRNA expression levels will require confirmation in larger cohorts. We also cannot exclude the possibility that specific characteristics of our study population, i.e. the strong negative correlation between age and BMI, may have produced a spurious relationship between FTO gene expression and obesity parameters. However, the fact that the BMI association withstood all conceivable adjustment makes us confident that this finding is robust. Adipose FTO expression and obesity could be coregulated by a common genetic or environmental factor and not functionally linked to each other. More sophisticated study designs are necessary to dissect causal from correlative relationships with FTO expression in fat.

We did not find an association between FTO mRNA levels and either fasting plasma glucose concentrations, HbA1c or parameters of insulin sensitivity. However, we cannot rule out the possibility that our sample size was too small to detect significant relationships between FTO gene expression and measures of insulin sensitivity and glucose metabolism. At the least, our data suggest that adipose tissue FTO expression does not directly affect risk of type 2 diabetes. As such, our results are generally supportive of the findings by Frayling et al. [3], who reported that the original association between FTO SNPs and type 2 diabetes disappeared upon adjustment for BMI.

Although expression of RPGRIP1L mRNA was also fat depot-specific in our study, it does not, in contrast to FTO, seem to be linked to the pathophysiology of obesity as no correlations with anthropometric and metabolic characteristics were found. Our data therefore favour FTO rather than RPGRIP1L as a possible target affected by reported obesity-associated polymorphisms [3]. Recently, mutations within RPGRIP1L have been shown to be responsible for several syndromes such as Joubert or Meckel syndrome [17, 18]. Joubert syndrome is an autosomal recessive multisystem disease characterised by cerebellar ataxia, developmental delay, hypotonia, irregular breathing pattern, eye movement abnormalities and cerebellar vermis hypoplasia and dysplasia with accompanying brainstem defects [18]. Meckel syndrome is an autosomal recessive lethal condition characterised by central nervous system malformation, postaxial polydactyly, cystic kidney dysplasia and ductal proliferation in the portal area of the liver [17, 18]. No obesity phenotype has been reported for these syndromes caused by RPGRIP1L mutations.

Given the strength of the correlation between FTO expression and adiposity in the present study and the relatively small effect that variation in FTO has on BMI as reported by Frayling et al. [3], variation in the FTO genotype does not appear to play a major role in regulating FTO transcription in adipose tissue. This is supported by the fact that no association between the obesity risk polymorphism and FTO and/or RPGRIP1L mRNA expression in adipose tissue was found in our study. It is noteworthy that, despite our study’s lack of statistical power due to the small sample size, rs8050136 was still associated with obesity in this study group.

In conclusion, expression of adipose tissue FTO mRNA is fat depot-specific and correlates with age, while being negatively correlated with BMI and total body fat. Adipose tissue FTO mRNA expression is not related to rs8050136 genotype, which is in LD with the previously reported obesity-associated SNP rs9939609 [3]. However, we were unable to determine the causal direction of the association between obesity and alterations in FTO gene expression. Therefore further studies are necessary to elucidate whether changes in adipose tissue FTO gene expression are secondary to obesity or causally related.