Emerging Function of Fat Mass and Obesity-Associated Protein (Fto)

Genome-wide association studies (GWAS) are a laborious but powerful tool to identify genetic risk factors associated with complex polygenic traits such as obesity [1], diabetes [2], or coronary artery disease [3]. The link between genetic variation in FTO and obesity was first described in a GWAS for type 2 diabetes [1] and was later independently confirmed in different populations all over the world. First described in 2007, genetic variation in FTO has since become one of the most solidly confirmed risk factor for polygenic obesity in humans; yet, information about how FTO affects metabolism is still scarce. Bioinformatic analyses suggest FTO codes for a Fe(II)and 2oxoglutarate–dependent nucleic acid demethylase [4,5] that catalyzes demethylation of 3-methylthymine in single-stranded DNA [5]. However, how this proposed function of FTO is integrated into the complex network of energy metabolism control remains the object of intense scientific investigation. Analyses of genetically engineered mouse models, in which the function of Fto is either eliminated [6–8] or enhanced [9], support a role of Fto in energy metabolism but are inconsistent as to whether Fto modulates caloric intake, energy expenditure, or both. In 2009, global lack of Fto was reported to result in a lean phenotype as a consequence of increased energy expenditure [7]. Similar results were reported by another study [8], but both studies share two pitfalls. First, interpreting energy expenditure differences can be challenging when body composition differences are also present (see below). Second, germline loss of Fto causes perinatal lethality and growth retardation, which may give rise to secondary effects that are unrelated to the mechanism by which variation in FTO affects human metabolism [7,8]. Notably, homozygous mice carrying a loss-of-function missense mutation in the C-terminal domain of Fto (367F) show no signs of perinatal lethality or growth retardation; they are lean, exhibit normal food intake and, when normalized by body weight, show increased energy expenditure [6]. While these studies appear to point to a potential role of Fto in regulating energy expenditure rather than food intake, mice globally overexpressing Fto are obese, hyperphagic, and exhibit normal energy expenditure when corrected for body fat or lean tissue mass [9]. In line with these data, most human studies report that obesity-predisposing FTO alleles are associated with increased food intake, but not energy expenditure (Table 1) [10–13]. In summary, despite intense scientific discussion about whether Fto primarily affects energy expenditure or food intake [14], it remains unclear what role Fto plays in early development compared to adult life and which tissues and/or brain regions are involved in mediating the effects seen in the global Fto knock-out (ko) mice. An important step in solving these questions has now been taken by Roger Cox and colleagues. In the current issue of PLOS Genetics, McMurray et al. [15] report a series of elegant studies further elucidating the complexity of Fto with respect to how, when, and where it is most relevant for energy metabolism and shedding new light on the recently proposed role of Fto in protein metabolism [16]. In their manuscript, the authors recapitulate that germline loss of Fto leads to perinatal lethality, growth retardation, and a lower body weight that is accompanied by decreased body fat and lean tissue mass. However, in contrast to previous reports, the authors convincingly show that there is no difference in energy expenditure when the data are interpreted correctly, i.e., using a regression approach that takes into account potential confounding by differences in lean body mass. Several review articles have recently highlighted this issue [17,18], and it is now clear that simply dividing raw energy expenditure results by lean body mass can lead to spurious conclusions. (This is illustrated quite nicely in Figure 3 of McMurray et al., where an apparent difference of energy expenditure per gram of lean body mass vanishes upon regression adjustment using ANCOVA.) Interestingly, total food intake was not changed in the germline Fto ko mice, whereas the ratio between CO2 production and O2 consumption (respiratory exchange ratio; RER) was decreased, suggesting that Fto ablation promotes protein and/or fat utilization. To circumvent the challenge of perinatal lethality and growth retardation, McMurray et al. used tamoxifen-inducible ubiquitin-cre mice to delete Fto at the time of sexual maturity. These adult onset Fto ko mice showed no increased lethality or growth retardation but had a lower body weight accompanied by a decreased lean mass and, interestingly, an increased body fat mass. No changes were observed in energy expenditure or total food intake, but, similar to the germline Fto ko, RER decreased in adult onset Fto ko mice, an effect also noted in a recent human study [19]. As central nervous system (CNS)-specific Fto deletion was recently reported to recapitulate the phenotype of the germline Fto ko mice [8], McMurray et al. further used an adenoviral associated approach to specifically knock out Fto in the mediobasal hypothalamus (MBH). Interestingly,