FTO gains function

Abstract

Previous genome-wide association studies have identified a strong association between FTO and human obesity, although the mechanism by which FTO affects obesity remains unknown. Anew study suggests that the obesity risk alleles are gain-of-function.

In 2007, a genome-wide association study identified the strongest association signal yet reported for an adiposity modifier gene¹, located in the first intron of FTO (also known as the fat mass and obesity–associated gene) (Fig. 1). Individuals homozygous for the FTO risk allele eat more² and are, on average, ~7 pounds heavier1 than individuals homozygous for the protective allele. Individuals heterozygous for the risk allele are 3 pounds heavier on average, implying a dose-dependent effect1. Although FTO was found to have demethylase activity³, this function provided no immediate clues as to the molecular-physiological mechanism by which this gene might affect adiposity.

Figure 1.

Genomic organization of the FTO locus on chromosome 16 (16q12.2, at 52.3–52.4 Mbp). rs9939609, in linkage disequilibrium (LD) with rs8050136, is associated with body mass index (BMI) and is located in a 47-kb LD block. FTO is an AlkB-like demethylase with a strong preference for 3-methylthymidine (3-metT) in single-stranded nucleic acids. RPGRIP1L is a component of the basal body of the cilium. The leptin (LEPR) and melanin-concentrating hormone 1 (MCHR1) receptors localize to neuronal cilia. CUX1, which binds at rs8050136, regulates FTO and RPGRIP1L expression from a site located in the region of association in intron 1 of FTO. Not drawn to scale. Black boxes indicate exons

Some studies have suggested that FTO exerts anorexigenic effects, as fasting (accompanied by an increased drive to eat) is associated with decreased hypothalamic Fto expression in mouse^3,4. However, mice segregating a null Fto allele are lean due to increased energy expenditure⁵. These mice also have higher postnatal mortality rates, generalized reduction in body size, hyperphagia and reduced body fat, suggesting that Fto may play a direct role in the development of adipose tissue. In contrast, mice homozygous for a dominant-negative Fto mutation that does not cause increased perinatal mortality or generalized reduction in body size display reduced fat mass, as well as increased energy expenditure without apparent changes in physical activity or food intake⁶. From these studies, it is not clear whether the susceptibility alleles in humans are hypermorphic or hypomorphic with regard to FTO function, whether the impact on obesity is conveyed entirely by effects in the brain or whether the effects on FTO expression differ by organ. On page 1086 of this issue, Roger Cox and colleagues report the first in vivo study that directly addresses the effects of systemic Fto overexpression on food intake and energy expenditure7.

FTO gain of function

Church et al.7 created mice carrying one or two extra copies of a ubiquitously expressed Fto gene, resulting in Fto overexpression in all tissues tested, including white adipose tissue (overexpressed by ~1.5-fold and ~3-fold, respectively) and hypothalamus (overexpressed by ~2-fold and ~2.5-fold, respectively), with the highest levels of overexpression seen in muscle (~3-fold and ~10-fold, respectively). These mice displayed increased energy intake and increased adiposity in a dose-dependent manner, in agreement with the FTO dose-dependent differences in adiposity seen in humans1. The authors also confirmed that Fto-null mice have reduced fat mass, as has been previously reported⁵. Taken together, these data suggest that increased FTO expression results in increased food intake, leading to increased adiposity. Thus, a gain-of-function effect is suggested for the implicated human allele.

Brain versus fat

Because FTO is ubiquitously expressed, it is not immediately apparent which organs are the primary mediators of the gene’s effects on adiposity. FTO has been implicated in the regulation of lipolysis⁸. A number of contradictory reports found that FTO expression in human adipose tissue is negatively correlated or uncorrelated with FTO susceptibility alleles^9,10. Moreover, FTO expression is higher in the subcutaneous fat of obese individuals⁸ independent of FTO genotype status. Relevant to a possible primary role of FTO in adipose tissue, Church et al.⁷ now report that fasted mice overexpressing Fto have lower circulating concentrations of leptin per unit of fat mass than fasted control animals. Leptin is synthesized in fat and provides a signal to the hypothalamus and other parts of the brain regarding the size of fat stores. Absolute or relative leptin deficiency results in hyperphagia¹¹. Thus, it is possible that systemic FTO overexpression affects leptin expression or secretion from fat and thereby affects central nervous system– mediated control of food intake.

A number of reports have also implied a direct role for FTO in the central nervous system in mediating food intake^3,4. Overexpression of Fto in the arcuate nucleus, a hypothalamic region that plays a role in regulating food intake, decreases energy intake¹², an effect opposite to that reported by Church et al.⁷ in mice overexpressing Fto systemically.

The data described above suggest opposing effects of FTO upregulation on energy intake in fat versus brain that cannot be resolved using a transgenic animal with generalized FTO overexpression. The generalized extreme overexpressor model may not reflect the molecular physiology of obesity in humans segregating for FTO alleles that predispose to obesity; it is possible, for example, that the functional variant(s) in humans result in subtly reduced FTO expression in the brain and subtly increased expression in fat or other tissues.

What about RPGRIP1L?

The SNPs associated with adiposity are part of a linkage disequilibrium block that includes parts of FTO intron 1, exon 2 and part of intron 2 (Fig. 1). Because no functionally significant sequence variants have been found in exon 2, it seems likely that the functional SNP(s) alter a regulatory element that controls transcription. RPGRIP1L, the retinitis pigmentosa GTPase regulator-interacting protein-1 like gene, is located ~100 bp 5′ _of and in the opposite transcriptional orientation as FTO and may also account for some or all of the region’s association with adiposity. RPGRIP1L is a component of the basal body of the primary cilium and might participate in human energy homeostasis by virtue of the obesity phenotypes in Bardet-Biedl and Alström syndromes, in which all implicated genes encode components of the primary cilium¹³. The leptin (LEPR) and melanin-concentrating hormone 1 (MCHR1) receptors (both involved in the regulation of feeding and energy balance) localize in the vicinity of neuronal cilia^4,14,15, which may affect leptin and/or MCH signaling. Hypothalamic Rpgrip1l expression, as with Fto, is decreased with fasting⁴, consistent with a role in energy homeostasis. In addition, FTO and RPGRIP1L are co-regulated by CUX1 from a site within the DNA region associated with obesity (Fig. 1)⁴. Therefore, the function of both FTO and RPGRIP1L may be affected in individuals carrying the obesity risk allele, which may also partly explain the strength of the association.

Thus, the molecular basis for the association of SNPs in intron 1 of FTO with adiposity in humans remains unknown. It is plausible that the locus contains elements that mediate the regulation of both FTO and RPGRIP1L (and, possibly, other more distant genes) and that this regulation involves the hypothalamus, adipocytes and possibly other cell types involved in the regulation of body weight. Tissue-specific underexpression and overexpression of both genes will be required to work out the complex molecular physiology of FTO-associated obesity. The effort is likely to be rewarded by identification of new regulators of body weight.