| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Neuroscience (R.F., M.H., P.K.O., O.S., J.A.J., A.M.O. J.L., H.B.S.), Functional Pharmacology, Uppsala University, Biomedical Center, SE 75124 Uppsala, Sweden; and Minnesota Obesity Center (P.K.O., A.S.L.), Department of Food Science and Nutrition (A.S.L.), St. Paul, Minnesota 55108
Address all correspondence and requests for reprints to: Robert Fredriksson, Department of Neuroscience, Biomedical Center, Box 593, SE 75124 Uppsala, Sweden. E-mail: robert.fredriksson{at}bmc.uu.se.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There are several genes associated with obesity on the human obesity map (1) such as the melanocortin 4 receptor (MC4R), leptin and the leptin receptor. However, the contribution of each specific gene to obesity is low, being highest for the MC4R gene ranging from 1–6% (2). The overall inheritability of BMI is estimated to be about 50–60%. Recently, three reports have shown a strong association of a single-nucleotide polymorphism (SNP) in a gene called FTO with both childhood and adult obesity. Frayling and colleagues (3) performed a genome-wide association study for about 490,000 autosomal SNPs in a type 2 diabetes population in the United Kingdom. They found that SNP rs9939609 in the FTO gene was strongly associated with type 2 diabetes, but this allele was also strongly associated with an increased BMI. The association between this FTO SNP and type 2 diabetes was abolished by adjustment for the BMI, suggesting that it was due to the increased BMI. The association of this variant with the BMI was replicated in 13 cohorts with over 38,000 individuals. Interestingly, 16% of the adults who were homozygous for this SNP weighed about 3 kg more and had 1.67-fold increased odds of obesity. This association was observed from age 7 yr upward, and it reflects a specific increase in fat mass (3). Independently, Dina et al. (4) found another SNP, rs1121980, in the first intron of the FTO gene, that was strongly associated with severe (BMI > 40) adult obesity with odds ratio of 1.55 in a population of French individuals of European ancestry). Further genotyping showed a similarly strong association of several SNPs in a cohort of about 900 severely obese adults and 2700 nonobese French controls. Three of the four most significantly associated SNPs (rs17817449, rs3751812, and rs1421085) are putatively functional, and they were found to be similar in males and females. This report also shows associations of SNPs in the FTO gene to obesity in three additional cohorts including either children or adults (4). In an additional independent study, Scuteri and colleagues (5) showed that another SNP in FTO, rs9930506, and other nearby variants are associated with BMI, hip circumference, and total body weight in a study including over 4000 Sardinians. Additional associations of the BMI and SNPs in the FTO gene were shown in both European-American and Hispanic-American cohorts in the same report.
The analysis of previous studies (5) shows that the FTO gene maps to a region where linkage to the BMI has been reported in two previous genome-wide linkage scans (LOD = 3.2) in the Framingham Heart Study (6), and LOD = 2.2 in families with white ancestry from the Family Blood Pressure Program (7). Moreover, a deletion of the chromosome region harboring the FTO gene in one single individual was shown to result in mental retardation, finger anomalies, and obesity (8). The FTO gene was originally described in the mouse with fused toes, which has deletion of at least six genes and gross developmental phenotype (9). The association with human obesity has prompted the HUGO Gene Nomenclature Committee to change the name to fat mass and obesity associated (FTO). The FTO is currently also found under the name Fatso at the National Center for Biotechnology Information (NCBI) websites.
Together these studies provide by far the most convincing evidence for a single gene variation that alters the BMI (3, 10). The association to obesity is very strong and replicable in most populations, but the mechanism by which variants in FTO lead to obesity are unknown. After submission of this manuscript, it was reported that the FTO gene encodes a 2-oxoglutarate-dependent nucleic acid demethylase (11). The functional role of the gene in association of regulation of food intake is, however, not well characterized in terms of anatomy or functional associations to the neuropeptides that play a key role in the regulation of energy homeostasis. Several molecular components of the central regulation of body weight are evolutionarily well conserved, whereas others are highly different between species (12, 13).
Here we studied the evolutionary origin of the FTO gene and examined FTO’s functional importance by determining its expression in different behavioral models of energy deficit, including a study of coregulation of FTO with other feeding related genes. We also performed detailed anatomical expression charting with emphasis on the central feeding circuits.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
These sequences were aligned using ClustalW 1.83 (15). Sequence Hidden Markov Models (HMMs) were constructed from alignments using HMMbuild from the HMMER (16) 2.3.2 package and calibrated using HMMcalibrate. Protein datasets (peptide all collections) were downloaded from ftp://ftp.ensembl.org/pub for the following species: Aedes aegypti (AegL1.46), Anopheles gambiae (AgamP3.46), Caenorhabditis elegans (WB170.46), Ciona intestinalis (JGI2.46), Ciona savignyi (CSAV2.0.46), and Drosophila melanogaster (BDGP4.3.46). The protein datasets were searched through with the FTO sequence HMM using HMMsearch, and all hits with E value less than 10 were manually investigated.
Phylogenetic analysis.
The sequences were manually edited, and exons unique for one species (most likely wrongly predicted exons from the gene prediction programs) were removed. Also, the highly variable N termini were excluded. Alignments were performed using ClustalW 1.83 (15) and manually edited to the alignment presented in Fig. 1
. The alignment was bootstrapped 1000 times using SEQBOOT from the Win32 version of the Phylip 3.6 package (17). Maximum-parsimony and neighbor joining trees were calculated on the bootstrapped alignment with PROTPARS and PROTDIST/NEIGHBOR from the Win32 version of the Phylip 3.6 package. All trees were unrooted. The parsimony trees were calculated using ordinary parsimony, and the topologies were obtained using the built-in tree-search procedure. Protein distances for the neighbor joining trees were calculated using the JTT amino acid model. Majority-rule consensus trees were constructed using CONSENSE from the Phylip 3.5 package. The trees were plotted using TreeView (18) and manually edited in Canvas 8 (ACD, Systems, Miami, FL).
|
3 mm thick) using the brain matrix as schematically shown in supplemental Fig. 1 (published as supplemental data on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org). The remaining brains were dissected in more detail using the rat brain matrix (coronal sections). The following regions were dissected from the sections with the guidance of a rat brain atlas (20): prefrontal cortex, cortex, and hippocampus. The first section (approximately bregma 5.20–4.20 mm) was denoted prefrontal cortex. The cortex was collected from sections approximately bregma 4.20–2.70 mm and included the primary and secondary motor cortex. The hippocampus was collected from several sections and therefore comprised the hippocampus CA1–3 and the dentate gyrus. The tissues were immersed in the RNAlater solution (Ambion, Austin, TX), kept at room temperature for 1 h, and thereafter stored at –80 C until further processed. The same animals as in Lindblom et al. (21) were used in this study. Briefly, 24 (eight animals per group) outbred male Sprague Dawley rats (Alab, Sollentuna, Sweden) with the initial body weight of about 225 g were randomized into three groups: control (ad libitum fed), food-restricted, and food-deprived animals. Animals in the control group had free access to R36 food pellets (Labfor, Lactamin), whereas the food-restricted group for 12 d received 45 ± 1% of the amount of food consumed by controls in the same time frame. During this period, the control animals gained 40 ± 1.5% in body weight, whereas the food-restricted animals lost 8 ± 0.9%. The food-deprived animals had access to food until the last 48 h of the experiment, when they were completely deprived of food. The animals were killed by decapitation between 3 and 6 h into the light period, and the brains were rapidly removed. At decapitation, blood was collected in EDTA-prepared tubes, spun for the preparation of plasma, which was later stored at –20 C until used for hormonal measurements (leptin, insulin, corticosterone, ACTH, ghrelin, and adiponectin). The entire hypothalamus was isolated using the brain matrix (Activational System) and the brain atlas as a guide (20). The rostral border was at the crossing of the anterior commissure (bregma approximately –0.4 mm) and the caudal border at the end of the mammillary recess of the third ventricle (bregma approximately –4.5 mm). The dissected tissue bordered dorsally to the bed nucleus of the stria terminalis (rostral part) and zona incerta (caudal part) and laterally to the substantia innominata (rostral part) and the optic tract and internal capsule (caudal part). Individual tissue samples were rapidly frozen on dry ice, immersed in RNAlater solution (Ambion, Stockholm, Sweden), kept at room temperature for approximately 1 h to allow the solution to infiltrate the tissue, and then stored at –80 C until further processed.
RNA isolation and cDNA synthesis.
Individual tissue samples were homogenized by sonication in the TRIzol reagent (Invitrogen, Breda, The Netherlands) using a Branson sonifier. Chloroform was added to the homogenate, which was then centrifuged at 10,000 x g at 4 C for 15 min. The water phase was transferred to a new tube, and RNA was precipitated with isopropanol. The pellets were washed with 75% ethanol, air dried at room temperature, and dissolved in RNase-free water. DNA contamination was removed by treatment with DNase I (Roche Diagnostics, Uppsala, Sweden) for 4 h at 37 C, and the DNase I was thereafter inactivated by heating the samples at 75 C for 15 min. The absence of genomic DNA was confirmed by PCR with primers for the rat and mouse RNA extractions with glyceraldehyde-3-phosphate dehydrogenase (GAPDH; NM_017008; forward TCC CTC AAG ATT GTC AGC AA, and reverse, CAC CAC CTT CTT GAT GTC ATC) on the Dnase-treated RNA. RNA concentration was determined using a Nanodrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE). cDNA was synthesized with MMLV reverse transcriptase (General Electric, Uppsala, Sweden), using random hexamers as primers according to the manufacturer’s instructions.
Real-time PCR.
The cDNA for the tissue panels was analyzed in quantitative real-time PCR with a MyiQ thermal cycler (Bio-Rad Laboratories, Stockholm, Sweden). Each real-time PCR with a total volume of 20 µl contained cDNA synthesized from 25 ng total RNA, 0.25 M each primer, 20 mM Tris/HCl (pH 8.4), 50 mM KCl, 4 mM MgCl2, 0.2 mM dNTP, SYBR Green (1:50,000). Real-time PCR was performed with 0.02 U/liter Taq DNA polymerase (Invitrogen) under the following conditions: initial denaturation for 3 min at 95 C, followed by 50 cycles of 15 sec at 95 C, 15 sec at 54–61 C (optimal annealing temperature), and 30 sec at 72 C. This was followed by 84 cycles of 10 sec at 55 C (increased by 0.5 C per cycle). All real-time PCR experiments were performed in duplicates, and the measurements where the threshold cycle (Ct) values between the duplicates had a difference of over 0.9 were repeated. A negative control for each primer pair and a positive control with 25 ng rat genomic DNA were included on each plate. For the food-restricted and food-deprived groups, RNA was extracted and converted into cDNA as described above, and quantitative real-time PCR was run as above, with the exception that six housekeeping genes were used (supplemental Table 1). The primer sequences for the rat FTO gene is also found in supplemental Table 1. Melting point curves were included after the thermocycling to confirm that only one product with the expected melting point was formed.
Data analysis and relative expression calculations.
The MyiQ software version 1.04 (Bio-Rad) was used to analyze real-time PCR data and derive Ct values. Melting curves were analyzed manually for each individual sample to confirm that only one product was amplified and that it was significantly shifted compared with the melting curve for the negative control. The sample Ct values were analyzed further if the difference between those and the negative control was greater than 2; otherwise, the transcript was considered not to be expressed. LinRegPCR (22) was used to calculate PCR efficiencies for each sample. After that, Grubbs’ test (GraphPad, San Diego, CA) was applied to exclude outliers and calculate average PCR efficiency for each primer pair. The 
Ct method (23) was used to transform Ct values into relative quantities with SD, and the highest expression for each primer pair was set to 1. Hereafter, the statistical analysis for tissue panels and treatment groups was conducted differently. For the tissue panels, all values in each data set were divided by the relative quantity for genomic DNA. The GeNorm software (23) was used on the two housekeeping genes to calculate normalization factors for every tissue to compensate for differences in cDNA amount. Thereafter, the normalized quantities were calculated and compared with genomic DNA, which was set to 100%. For the food-restricted and food-deprived groups, the analysis of real-time PCR data were performed as we have previously reported (21). The data were corrected for primer efficiencies and normalized as above. Differences in gene expression between groups were analyzed using ANOVA followed by Tukey’s post hoc test on genes significantly up- or down-regulated in the ANOVA. P < 0.05 was used as the criterion of statistical significance for the ANOVA. Statistics were performed using Prism (GraphPad).
In situ hybridization and immunohistochemistry
Animals.
Adult male wild-type Sv129 mice were anesthetized, and the brains were fixed by transcardial perfusion before being excised. For paraffin-embedded tissue sections, the brains were fixed in zinc-formalin (Richard-Allan Scientific, Kalamazoo, MI) for 18–24 h at 40 C before dehydration and paraffin infusion (Tissue-Tek vacuum infiltration processor; Miles Scientific, Naperville, IL). Sections were cut at 7 µm using a Microm 355S STS cool-cut microtome, placed on Superfrost Plus slides (Menzel-Gläser, Braunschweig, Germany), incubated at 37 C for 18–24 h and stored at 4 C. For free-floating tissue sections, the brains were cut at 50 µm using a Leica VT1000S vibratome. For details see Lagerstrom et al. (24).
Synthesis of RNA probes.
The FTO cDNA clone (partial expressed sequence tag clone) was obtained from Invitrogen (Groningen, The Netherlands; Invitrogen clone ID 5150225) as a glycerol stock. Plasmid DNA preparation was performed using JETstar 2.0 Plasmid Purification Midi Kit/50 (Genomed, Lohne, Germany). The clone was sequenced at MWG (Ebersberg, Germany) (https://ecom.mwgdna.com) and confirmed to contain 773 bp of the 3'-untranslated region of the mouse FTO transcript. Twenty micrograms of plasmid DNA were linearized by digestion with 30 U SpeI (Fermentas, Vilnius, Lithuania) for 3 h. Ribo probes were synthesized using 1 µg of the template, RiboLock RNase inhibitor (Fermentas, Helsingborg, Sweden) and 40 U RNA polymerase in the presence of digoxigenin-11-UTP or fluorescein-12-UTP (Roche Diagnostics). Probes were quantified using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies).
In situ hybridization combined with immunohistochemistry.
Paraffin-embedded tissue sections were deparaffinized by incubation in X-tra Solve (Medite Histotechnik, Burgdorf, Germany) and rehydrated by successive washes in ethanol/diethylpyrocarbonate (DEPC), ending up in 100% DEPC. Sections were fixed for 10 min in 4% paraformaldehyde followed by washes in PBS before digestion for 15 min in 20 µg/ml proteinase K (Sigma-Aldrich, Stockholm, Sweden) diluted in 10 mM Tris-HCl (pH 8.0). The sections were refixed in 4% paraformaldehyde, and additional washes in PBS were performed. Acetylation treatment for 10 min in a mixture of 1.3% triethanolamine (Sigma-Aldrich), 0.2% acetic anhydride (Sigma-Aldrich), and 0.06% HCl diluted in DEPC was done, and the sections were then permeabilized in 1% Triton X-100 (Sigma-Aldrich) in PBS for 30 min. The sections were rinsed in PBS and placed in a humidified chamber [50% formamide, 5x standard saline citrate (SSC)] before preincubation for 2–5 h in hybridization buffer. Hybridization buffer consisted of 50% formamide, 20x SSC, 50x Denhardt's solution (1% Ficoll 400, 1% polyvinylpyrrolidone and 1% BSA; Sigma-Aldrich), 10 mg/ml yeast tRNA (R6750; Sigma-Aldrich), and 10 mg/ml sheared salmon sperm single-stranded DNA (Ambion, Sweden) in 0.1% DEPC. The final concentration of 1 µg probe/ml was then heat denatured in hybridization buffer and added onto the slides. The sections were covered with glass coverslips, and hybridization was performed for 16 h at 55 C. Prewarmed 5x SSC was used to remove coverslips, and the sections were incubated in prewarmed 0.2x SSC for 1 h at 55 C and for an additional 5 min at room temperature. The sections were washed in Tris-buffered saline (TBS) and placed in a humidified chamber containing TBS. Preincubation was performed in the blocking solution (1% blocking reagent; Roche Diagnostics) in TBS followed by incubation in alkaline phosphate-conjugated anti-fluorescein Fab fragments (1:5000; Roche Diagnostics) together with the primary antibody against the rabbit glial fibrillary acidic protein (GFAP; Sigma-Aldrich) diluted 1:320 in the blocking solution. The sections were incubated in the antibody solution overnight at 4 C. Sequential washes with 2 mM levamisole (GTF Fisher, Frolunda, Sweden) in TBS with 0.1% Tween 20 followed by washes with 2 mM levamisole in 100 mM NaCl, 10 mM Tris-HCl (pH 9.5), 50 mM MgCl2, and 0.1% Tween 20 were performed before color development of the alkaline phosphate-labeled probe with BM Purple or Fast Red enzyme substrate (Roche Diagnostics). The sections were washed in PBS and incubated with 4',6'-diamidino-2-phenylindole (DAPI; 1:1000; Roche Diagnostics) in DEPC for 5 min. An additional PBS rinse was performed before the sections were incubated with Alexa fluorescent secondary antibody (1:500; Invitrogen) for 2 h in the dark. After mounting in DTG mounting media with antifade [1.25 g (2.5%) DABCO (Sigma D-2522) in 45 ml glycerol and 5 ml 0.5 Tris (pH 8.6)], the sections were analyzed with the Olympus (Japan) BX61W1 microscope and the Optigrid system using the Volocity software (Improvision, Tubingen, Germany); the staining was also analyzed by confocal microscopy using the Zeiss (Oberkochen, Germany) LSM 510 META system.
In situ hybridization on free-floating sections
In situ hybridization was performed according to Lagerstrom et al. (24) with 500 ng/ml probe.
Ethical statement
All animal procedures were approved by the local ethical committee in Uppsala (permit C 275/4 and C 156/4) and followed the guidelines of the European Communities Council Directive (86/609/EEC).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and 2). Thirteen of these were of vertebrate origin, whereas two originated from two species of green algae, O. tauri and O. lucimarinus. We constructed a sequence Hidden Markov Model (HMM) from these sequences and searched the protein dataset of several invertebrate species: A. aegypti, As gambiae, C. elegans, D. melanogaster, C. intestinalis, C. savignyi, Branchiostoma floridae, and Nematostella vectensis. These species represent at least four distinct developmental lineages of animals, ecdysozoa (with nematoda and arthropoda), tunicates, cephalochordates, and cnidarians, and no FTO sequence could be found. We also searched the chicken genome at UCSC (http://genome.ucsc.edu/cgi-bin/hgBlat/) and identified a partial FTO sequence representing exon 1, part of exon 2, exons 3–4, and exon 6 (data not shown), although we were unable to identify the rest of the gene due to gaps in the genome assembly. In addition, we searched the 1.4x shotgun assembly of a cartilaginous fish, the elephant shark (at http://esharkgenome.imcb.a-star.edu.sg/) using translated blast with the human FTO as bait. We identified single-read sequences representing exons 2–4, 6, and 7 (not shown). These data suggest that FTO is present as a functional gene also in birds and cartilaginous fish, although we cannot fully exclude the possibility that these two sequences are pseudogenes. We did not identify FTO in prevertebrate chordates (B. floridae or Ciona), insects, C. elegans, Nematostella, yeast, or green plants. Taken together, FTO appears to be present within vertebrates, missing in invertebrate animals, fungi, and green plants, but present in algae, which split from the lineage leading to vertebrates early in eukaryotic evolution about 850 (25) to 1200 (26) million years ago.
|
Alignment and phylogeny
The amino acid sequence alignment of the FTO sequences is shown in Fig. 1. Gray boxes mark conserved residues, and black boxes mark splice sites. As for most genes, the exon-intron pattern is largely conserved within vertebrates, but interestingly, the algae sequences are intronless. In Fig. 1, we have also marked the part of the sequence (gray bar above the alignment) that can be aligned with AlkB from bacteria (11). At the global level, it can be seen that there is a large insertion of up to 85 residues in the first exon in FTO compared with AlkB. Interestingly, this region is also much smaller in algae compared with vertebrates, suggesting that this region expanded relatively recently in evolution. This region also displays a relatively low degree of conservation compared with the rest of the FTO gene, suggesting it could function as a spacer rather than as a functional domain. Also, the FTO sequences are significantly longer in the C-terminal region than AlkB, with the region corresponding to the last three exons in FTO being unique compared with AlkB. In Fig. 1, we have also marked four highly conserved regions in FTO, denoted I–IV (see supplemental Fig. 2). Interestingly, one of the most highly conserved regions lies in the part of FTO that is not present in AlkB, suggesting that this region could provide functionality or specificity that is unique for FTO compared with AlkB.
We made a phylogenetic tree using the maximum parsimony method (see Fig. 2) containing the predicted protein sequences of FTO from 16 species. The mammalian FTO sequences formed three clear clusters, one each for primates, rodents, and other mammals. The opossum sequences placed basal of the mammals as expected. The sequences from nonmammalian species formed three clusters, one for frogs, one for teleost fish, and one for the algae sequences. All the nodes separating these clusters have a bootstrap support over 95%. To investigate the evolutionary distance between the sequences, we mapped maximum likelihood branch lengths onto the maximum parsimony tree topology from Fig. 2 (cells below and to the left of the diagonal in Table 1). In Table 1, we also display pair-wise amino acid identity (nonshaded cells). We plotted the evolutionary time from humans (evolutionary time from hsa vs. hsa = 0) to the split of the other species, where the slope of the line represents evolutionary rate (data not shown). The evolutionary distance is proportional against evolutionary time for all animals, with the exception of Tetraodon, as indicated in Table 1. This shows that FTO in Tetraodon appears to be evolving much more rapidly than FTO in other species, or this evolutionary rate could also be due to the low-quality regions in the genomic sequence from which the Tetraodon sequence is derived. Interestingly, FTO in algae appears to be evolving much more slowly and is only twice that of the distance between FTO in human and fish.
|
. The results show that the FTO gene is expressed in all tested tissue types including all brain slices encompassing the entire rat brain.
|
and Fig. 4). Among the hypothalamic regions, particularly high density of labeling was observed in the arcuate (ARC), ventromedial (VMH), and supraoptic (SON) nuclei. FTO was distributed uniformly throughout the ARC, whereas the signal seemed more pronounced in the dorsal vs. ventral portion of the SON. The paraventricular nucleus (PVN) also contained FTO mRNA, primarily in the medial parvocellular and medial magnocellular subdivision. Sparse FTO-expressing cells were detected in the medial preoptic area (MPOA), yet they were scattered in the entire region. In addition, FTO mRNA was seen in the dorsomedial hypothalamic nucleus (DMH) and the periventricular zone around the third ventricle (Fig. 4). The analysis of extrahypothalamic sites revealed a high level of FTO expression in all portions of the bed nucleus of the stria terminalis (BNST). FTO mRNA was detected in the brainstem, including the area postrema (AP) and nucleus of the solitary tract (NTS). This gene was present only in the caudal part of the NTS. Aside from the expression in feeding-related areas, FTO mRNA was found in selected regions involved in visual processing and circadian rhythmicity, such as the suprachiasmatic nucleus (SCN) (Table 2 and Figure 4), but also dorsal lateral geniculate nucleus (DLGN), intrageniculate leaflet (IGN) and thalamic PVN (tPVN). Positive FTO signal was also seen in the hippocampus, cortex, and medial vestibular nucleus (not shown).
|
|
shows fluorescent in situ hybridization of the mouse cortex (left) and mouse thalamus (right) with FTO labeled in red. Green labels indicate immunohistochemical staining for GFAP, a marker for astrocytes and glia cells. Blue staining is DAPI staining for nuclear DNA and, hence, serves as a marker for all cells. It is evident that there are cells staining specifically for green and blue (arrows marked with asterisks) as well as red and blue (unmarked arrows), but we found only one cell that had FTO overlapping with GFAP. This indicates that FTO is exclusively neuronal, possibly with very few exceptions.
|
). Surprisingly, the effect of starvation on FTO mRNA levels in rats is opposite to that observed in mice even though a similar 48-h deprivation model was used (11). There could be several reasons for this discrepancy, but at this time, they are all speculative. It is well established that mice show much higher sensitivity to starvation compared with rats. For example, Lenaerts and colleagues (29) studied protein expression in the small intestine at several stages of food deprivation. They found that as little as a 12-h difference in the length of starvation triggers a number of differential down- and up-regulation mechanisms in mice. Another possible cause of the difference could stem from circadian rhythmicity. It should be noted that aside from the food intake network, we also found that FTO is widely expressed within the visual/circadian system, including the SCN, DLGN, IGN, and tPVN, which suggests a possible involvement in the processing of visual information (30). Noteworthy, Gerken and co-workers (11) killed the mice at the very onset of the light phase of the light-dark cycle, whereas the rats in our study were decapitated in the middle of the light phase. Because several feeding-related hypothalamic sites that express FTO, including the PVN and ARC, respond to circadian cues and receive direct projections from the SCN (31), it is possible that FTO expression in the hypothalamus can be modified by entrainment.
|
-MSH), prolactin-releasing peptide (PrRP), somatostatin, TRH, and vasopressin as well as circulating levels of leptin, insulin, corticosterone, ACTH, ghrelin, and adiponectin in the same rats using validated quantitative PCR and hormone measurements (Johansson, A., R. Freddriksson, S. Winnergren, M. C. Lagerstrom, A.-L. Hulting, H. B. Schioth, and J. Lindblom, submitted). This study shows, in line with large amount of literature, that plasma leptin and POMC expression are low, whereas, for example, the expression of NPY was significantly higher under both conditions as seen in Fig. 6. Here we measured the expression level of the FTO gene in the same manner in the same animals and correlated it to the expression levels of these neuropeptides and hormones. The FTO expression levels showed a strong negative correlation with expression levels of four neuropeptides. The P values were 0.0245, 0.0087, 0.0022, and 0.0007 against GHRH, dynorphin, galanin, and GALP, respectively. This analysis was a multivariate (all against all) of in total 28 parameters with Spearman correlations, and hence, the analysis involves 392 independent tests. If these data are corrected using the Bonferroni approach, which yields an adjusted P value for 5% significance of 0.00012, GALP is still borderline significant. This is accepted as a very strict correction for multiple testing for a large number of tests (32). We also identified two neuropeptides as positively correlated to FTO, TRH and NPY with P values of 0.01 and 0.03, respectively. Using the less strict, and possibly more appropriate method for a large number of comparisons (32), false discovery rate, expression of GALP is still significantly correlated with expression of FTO (0.03), whereas galanin is borderline significant (0.06) at the 5% confidence level. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
), we suggest that the most likely scenario is that the gene was transferred from a vertebrate, perhaps a fish, to the algae via an mRNA intermediate. It is hard to date these events. but because FTO is not present in land plants (rice and Arabidopsis), this transfer probably occurred after the divergence of green plants from algae, approximately 480 million years ago (33). From the identification of FTO in vertebrates, we know it was present at least before the split of sharks from the lineage leading to mammals but most likely not before the split of amphioxus from the lineage leading to mammals.
The overall sequence is well conserved with 45% amino acid identity between human and zebrafish, the most distant vertebrate orthologs to the human FTO (Table 1). Figure 1 and supplemental Fig. 2 show that there are four regions within the FTO gene that are particularly well conserved. These are likely to be part of the most important functional domains of the gene. Three of these are homologs to the ALK gene (11), whereas the fourth is unique to FTO. It is interesting that three of these four regions, including the one unique to FTO, are also highly conserved in the algae sequences (see further analysis in the legend to supplemental Fig. 3).
The whole-body tissue panel shows that the FTO gene has wide-ranging expression in both peripheral and central tissues in the rat, which is in line with what was reported earlier for human tissues (3, 11). This study confirms that in rodents, FTO is widely expressed in the periphery and in the brain and enriches the previous findings in the mouse by adding twice as many tissue types in the analysis. Similar to Gerken et al. (11), we detected this gene in the cortex, hypothalamus, and cerebellum; however, we also found it in other central areas, including the brainstem, olfactory bulbs, pituitary, hippocampus, and prefrontal cortex. Our in situ data complement the real-time PCR results and show that within the hindbrain, FTO mRNA is particularly abundant in the NTS and AP, which integrate a plethora of peripheral and central signals. In the forebrain, FTO is present in discrete areas implicated in the regulation of hunger- and reward-driven consummatory responses. The comparison between the rat and mouse tissue panels presented herein and in the study by Gerken and colleagues (11) does not reveal significant species-related discrepancies. The differences are rather associated with the amount of mRNA detected in various tissue types relative to each other. We observed high FTO expression in the brain; however, some peripheral organs, such as the eye and skeletal muscle, displayed equally high levels of FTO mRNA content.
Here we focused particularly on the central circuitry associated with the regulation of food intake. Our in situ expression data suggest that FTO’s presumed involvement in body weight control could stem from its activity within nuclei that govern feeding. FTO appears to be expressed in several brain areas within this network, and it is particularly abundant in specific nuclei of the brainstem and hypothalamic structures, such as the NTS, AP, ARC, DMH, VMH, PVN, and SON, which are the relay stations integrating both central and peripheral information pertaining to food intake. The AP, NTS, PVN, and SON respond to peripheral signals, such as changes in plasma osmolality, presence of toxins, or the level of stomach distension, whereas the ARC receives stomach ghrelin input and adipose tissue-derived leptin signals (34, 35, 36). FTO-positive cells in the vicinity of the third ventricle are likely capable of integrating peptidergic information mediated via the cerebrospinal fluid. The NTS hosts vagal afferents and transmits vagally mediated input to, among others, the PVN, SON, and ARC (30, 35, 36). The forebrain sites are reciprocally connected with each other and with the brainstem, forming a complex network. The double labeling showed that the FTO gene is predominantly expressed in neurons, whereas it is in principal absent in astrocytes or glia cells. Taken together, the data show an expression pattern that suggests that this gene could have a very specific role in the neuronal network regulating feeding.
These neuroanatomical data do not, however, allow one to link FTO with any particular aspect of feeding, such as hunger, satiety, preference, or aversion. Noteworthy, FTO mRNA is present not only in brain regions that are thought of as important for energy balance but also in regions thought of as reward related. Although the role of hypothalamic and brainstem sites seems to be associated to a large extent with maintaining the proper energy balance, the BNST, as part of the extended amygdala, affects rewarding aspects of consumption. One should note, however, that the distinction between sites influencing consumption for reward vs. energy is an oversimplification, because they tend to contain peptides and/or receptors involved in both types of ingestive behavior. Moreover, several of the sites where we find the FTO expression contain both anorexigenic and orexigenic peptides, such as ARC neurons that synthesize -MSH and CART to inhibit feeding as well as agout-related peptide (AgRP), β-endorphin, and NPY, which are hyperphagic (36).
The fact that FTO expression was significantly up-regulated in the hypothalamus of rats during 48 h of food deprivation provides further support that this gene could have a regulatory role in the central feeding circuits. Moreover, FTO expression levels showed correlation with some feeding-related genes that are present within the same brain circuitry as FTO itself, with strong correlation with GALP. The other correlations of NPY, galanin, dynorphin, GHRH, and TRH were not significant considering a strict accounting for multiple testing but provide clues for further studies. These peptides all participate in food intake regulation and may have overlapping expression with the FTO gene (see in situ data), and it is possible that they could also follow a specific pattern of expression dependent on the level of adiposity.
Although we have focused on the main food intake circuits in this study, we also show that the FTO gene has a specific expression pattern in other parts of the brain. For example, we find consistent expression within the visual/circadian system circuitry, including the SCN, DLGN, IGN, and tPVN, which suggests a possible involvement in the processing of visual information at the central level (36). There is also abundant presence of FTO in the hippocampus and cortex, and hence, it is likely that FTO, similar to many central nervous system genes, displays a pleiotropic character influencing several systems.
In summary, we have shown that an ancestor of the FTO gene was present at least 450 million yr ago. The FTO is likely to be present in most vertebrate species in a single copy. This gene is expressed in many tissues including the central nervous system and its feeding circuitry, where it is found in distinct nuclei of the hypothalamus, brainstem, and extended amygdala. FTO seems to be specific for neuronal cell populations. Expression of this gene is up-regulated during starvation, and it is correlated to the orexigenic peptide GALP, which is mainly expressed in the ARC. This study provides evidence supporting the central role of the FTO gene in the regulation of energy homeostasis, although it is clear that the widespread expression of FTO strongly indicates this gene’s involvement in other functions as well.
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online January 24, 2008
Abbreviations: AP, Area postrema; ARC, arcuate nucleus; BMI, body mass index; BNST, bed nucleus of the stria terminalis; Ct, threshold cycle; DAPI, 4',6'-diamidino-2-phenylindole; DEPC, diethylpyrocarbonate; DMH, dorsomedial hypothalamic nucleus; DLGN, dorsal lateral geniculate nucleus; GALP, galanin-like peptide; GFAP, glial fibrillary acidic protein; IGN, intrageniculate leaflet; MPOA, medial preoptic area; NPY, neuropeptide Y; NTS, nucleus of the solitary tract; 2OG-Fe(II), Fe(II)-dependent 2-oxyglutarate oxygenase; POMC, proopiomelanocortin; PVN, paraventricular nucleus; SCN, suprachiasmatic nucleus; SNP, single-nucleotide polymorphism; SON, supraoptic nucleus; SSC, standard saline citrate; TBS, Tris-buffered saline; tPVN, thalamic PVN; VMH, ventromedial nucleus.
Received October 23, 2007.
Accepted for publication January 14, 2008.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. S.F. Doney, J. Dannfald, C. H. Kimber, L. A. Donnelly, E. Pearson, A. D. Morris, and C. N.A. Palmer The FTO Gene Is Associated With an Atherogenic Lipid Profile and Myocardial Infarction in Patients With Type 2 Diabetes: A Genetics of Diabetes Audit and Research Study in Tayside Scotland (Go-DARTS) Study Circ Cardiovasc Genet, June 1, 2009; 2(3): 255 - 259. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Pausova, C. Syme, M. Abrahamowicz, Y. Xiao, G. T. Leonard, M. Perron, L. Richer, S. Veillette, G. D. Smith, O. Seda, et al. A Common Variant of the FTO Gene Is Associated With Not Only Increased Adiposity but Also Elevated Blood Pressure in French Canadians Circ Cardiovasc Genet, June 1, 2009; 2(3): 260 - 269. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hakanen, O. T. Raitakari, T. Lehtimaki, N. Peltonen, K. Pahkala, L. Sillanmaki, H. Lagstrom, J. Viikari, O. Simell, and T. Ronnemaa FTO Genotype Is Associated with Body Mass Index after the Age of Seven Years But Not with Energy Intake or Leisure-Time Physical Activity J. Clin. Endocrinol. Metab., April 1, 2009; 94(4): 1281 - 1287. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
