articles

genetics

nature publishing group

The FTO Gene Is Associated With Adulthood Obesity in the Mexican Population
Marisela Villalobos-Comparn1, M. Teresa Flores-Dorantes1, M. Teresa Villarreal-Molina1, Maricela Rodrguez-Cruz2, Ana C. Garca-Ulloa3, Lorena Robles3, Adriana Huertas-Vzquez4, Nubia Saucedo-Villarreal5, Mardia Lpez-Alarcn2, Fausto Snchez-Muoz6, Aarn Domnguez-Lpez6, Ruth Gutirrez-Aguilar7, Marta Menjivar8, Ramn Coral-Vzquez9, Gabriel Hernndez-Stengele9, Victor S. Vital-Reyes10, Victor Acua-Alonzo11, Sandra Romero-Hidalgo12, Doris G. Ruiz-Gmez3, Daniela Riao-Barros3, Miguel F. Herrera3, Francisco J. Gmez-Prez3, Philippe Froguel7, Eduardo Garca-Garca3, M. Teresa Tusi-Luna1, Carlos A. Aguilar-Salinas3 and Samuel Canizales-Quinteros1
Common polymorphisms in the fat mass and obesity-associated gene (FTO) have shown strong association with obesity in several populations. In the present study, we explored the association of FTO gene polymorphisms with obesity and other biochemical parameters in the Mexican population. We also assessed FTO gene expression levels in adipose tissue of obese and nonobese individuals. The study comprised 788 unrelated Mexican-Mestizo individuals and 31 subcutaneous fat tissue biopsies from lean and obese women. FTO single-nucleotide polymorphisms (SNPs) rs9939609, rs1421085, and rs17817449 were associated with obesity, particularly with class III obesity, under both additive and dominant models (P = 0.0000004 and 0.000008, respectively). These associations remained significant after adjusting for admixture (P = 0.000003 and 0.00009, respectively). Moreover, risk alleles showed a nominal association with lower insulin levels and homeostasis model assessment of B-cell function (HOMA-B), and with higher homeostasis model assessment of insulin sensitivity (HOMA-S) only in nonobese individuals (Pdom = 0.031, 0.023, and 0.049, respectively). FTO mRNA levels were significantly higher in subcutaneous fat tissue of class III obese individuals than in lean individuals (P = 0.043). Risk alleles were significantly associated with higher FTO expression in the class III obesity group (P = 0.047). In conclusion, FTO is a major risk factor for obesity (particularly class III) in the MexicanMestizo population, and is upregulated in subcutaneous fat tissue of obese individuals.
Obesity (2008) 16, 22962301. doi:10.1038/oby.2008.367

Introduction

The rise in the prevalence of obesity has been described as a worldwide increasing health problem and a risk factor for various disorders including type 2 diabetes (T2D) and cardiovascular disease (1). In a recent genome-wide scan association study comparing T2D patients and control individuals, a common polymorphism (rs9939609) in the fat mass and obesityassociated gene (FTO) showed strong association with T2D
1

and obesity; however, the association with T2D was abolished after adjustment for BMI (2). Dina et al. (3) concurrently demonstrated the contribution of FTO variants to childhood and class III adult obesity in the French population, and several studies have accordingly reported the association of FTO gene polymorphisms with obesity, mainly in white populations (48). The strong replication of association with obesity for FTO variants in different European populations make this gene

Unit of Molecular Biology and Genomic Medicine, Instituto Nacional de Ciencias Mdicas y Nutricin Salvador Zubirn, Instituto de Investigaciones Biomdicas de la Universidad Nacional Autnoma de Mxico, Mexico City, Mexico; 2Unit of Medical Research in Nutrition, Hospital de Pediatra, Centro Mdico Nacional Siglo XXI, Instituto Mexicano del Seguro Social, Mexico City, Mexico; 3Department of Endocrinology and Metabolism, Instituto Nacional de Ciencias Mdicas y Nutricin Salvador Zubirn, Mexico City, Mexico; 4Department of Human Genetics, David Geffen School of Medicine at UCLA, University of California Los Angeles, Los Angeles, California, USA; 5Faculty of Chemical and Biological Science, Universidad Autnoma de Sinaloa, Culiacn, Mexico; 6Department of Gastroenterology, Instituto Nacional de Ciencias Mdicas y Nutricin Salvador Zubirn, Mexico City, Mexico; 7Centre National de la Recherche Scientifique 8090, Institute of Biology, Pasteur Institute, Lille, France; 8Department of Biology, Facultad de Qumica, Universidad Nacional Autnoma de Mxico, Mexico City, Mexico; 9Unit of Medical Research in Human Genetics, Hospital de Pediatra, Centro Mdico Nacional Siglo XXI, Instituto Mexicano del Seguro Social, Mexico City, Mexico; 10Department of Biology of Reproduction, Hospital de Ginecologa y Obstetricia # 3, Centro Mdico Nacional La Raza, Instituto Mexicano del Seguro Social, Mexico City, Mexico; 11Molecular Genetics Laboratory, Escuela Nacional de Antropologa e Historia, Instituto Nacional de Antropologia e Historia, Mexico City, Mexico; 12National Coordination of Genomic Medicine, Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado, Mexico City, Mexico. Correspondence: Samuel Canizales-Quinteros (cani@servidor.unam.mx) Received 10 February 2008; accepted 30 May 2008; published online 31 July 2008. doi:10.1038/oby.2008.367
2296 VOLUME 16 NUMBER 10 | OCTOBER 2008 | www.obesityjournal.org

articles
genetics
one of the most important loci implicated in polygenic obesity. However, the contribution of FTO common variants to obesity is apparently not relevant in African American and several Asian populations (Chinese Han and native Oceanic) (9,10). Therefore, it is important to investigate whether FTO variants are associated with obesity and T2D in other ethnicities. It is currently known that FTO encodes a 2-oxoglutaratedependent nucleic acid demethylase (11). The human FTO gene is widely expressed in both fetal and adult tissues including adipose tissue, with the highest relative levels found in the brain and pancreatic islets (2). However, the role of FTO in obesity remains to be elucidated. Thus, we explored the association of the several FTO gene polymorphisms with obesity, T2D, and other biochemical parameters in the Mexican population as well as gene expression levels in adipose tissue of obese and nonobese individuals.
Methods And Procedures The study comprised two groups of individuals. The first group included 788 unrelated Mexican Mestizos aged 1882 years (544 nonpregnant women and 244 men); 653 were recruited from five different governmental institutions in Mexico City (Centro Mdico Nacional Siglo XXI, Instituto Nacional de Ciencias Mdicas y Nutricin Salvador Zubirn (INCMNSZ), Instituto Nacional de Neurologa y Neurociencias Manuel Velasco Surez, Universidad Nacional Autnoma de Mxico, and Universidad Autnoma Metropolitana-Iztapalapa). Because only 13 of these individuals had class III obesity, we extended this group recruiting 135 additional class III obese patients from the obesity clinic of the INCMNSZ. Individuals were grouped according to BMI: 424 were nonobese (BMI< 30kg/m2), 216 had class I/II obesity (30 BMI < 40kg/m2), and 148 had class III obesity (BMI 40kg/m2). In addition, 243 of these individuals had been diagnosed with T2D according to WHO (World Health Organization) criteria (12). Only individuals born in Mexico whose parents and grandparents identified themselves as Mexican Mestizos were included. The second group included 112 nondiabetic native Amerindian individuals: 40 Yaquis from the state of Sonora located in northern Mexico, 24 Purpechas from the state of Michoacn in western Mexico, and 48 Zapotecs from the state of Oaxaca, in southeastern Mexico. Biochemical parameters were measured as previously described (13). Homeostasis model assessment of B-cell function (HOMA-B) and homeostasis model assessment of insulin sensitivity (HOMA-S) as measures of B-cell function and insulin sensitivity were estimated using a computer model (14). This project was approved by the Institutional Committee of Biomedical Research in Humans of the INCMNSZ. All Mexican-Mestizo participants provided written informed consent prior to their inclusion in the study. The local authorities of Amerindian populations gave their approval to participate in the study, and a translator was used as needed. Single-nucleotide polymorphism genotyping The FTO rs9939609, rs1421085, and rs17817449 polymorphisms were genotyped using TaqMan assays (ABI Prism 7900HT Sequence Detection System; Applied Biosystems, Foster City, CA). Genotyping call rate exceeded 95% per single-nucleotide polymorphism (SNP), and no discordant genotypes were observed in 25 duplicate samples. In addition, because the Mexican-Mestizo population is admixed, it was necessary to analyze ancestry informative markers to assess whether any association could be confounded by population stratification. A panel of 10 ancestry informative markers (rs4884, rs2695, rs17203, rs2862, rs3340, rs722098, rs203096, rs223830, rs1800498, and rs281478) distinguishing mainly Amerindian and European ancestry ( >0.44) was
obesity | VOLUME 16 NUMBER 10 | OCTOBER 2008

screened in the population studied (15,16). Genotyping was performed by KBiosciences (Hertfordshire, UK, http://www.kbioscience.co.uk/) using a KASPar assay system. Genotyping call rates of each ancestry informative marker exceeded 95%, and no discordant genotypes were observed in 54 duplicate samples. Deviation from HardyWeinberg equilibrium was not observed for any SNP.
Subcutaneous fat tissue biopsies Fat tissue biopsies from 31 nondiabetic women (10 lean (age 34.4 8.0 years, BMI 23.9 1.2kg/m2), 10 class I/II obese (age 43.1 8.7 years, BMI 32.6 2.5kg/m2), and 11 with class III obesity (age 37.6 6.3 years, BMI 44.1 6.4kg/m2)) were used to assess FTO gene expression. FTO genotypes were available from 24 of these individuals (nine lean, eight class I/II obese, and seven class III obese subjects). Tissue samples from class III obese individuals were obtained during bariatric surgery, and samples from lean and class I/II obese individuals were obtained under local anesthesia. Samples were frozen in liquid nitrogen immediately after the biopsy and stored at 70C until processing. Total RNA was extracted using RNeasy Lipid kit (QIAGEN Sciences, Germantown, MD). RNA integrity was confirmed on agarose gels, and RNA concentration was measured by absorbance. FTO mRNA gene expression analysis For this purpose, 500ng of total RNA was reverse transcribed with TaqMan Reverse Transcription Reagents (Applied Biosystems) using random hexamers, according to the protocol recommended by the manufacturer. Real-time PCR was performed in a LightCycler 2.0 (Roche, Rotkreuz, Switzerland), using LNA TaqMan probes from the Universal Probe Library (Roche), in combination with intron-spanning specific primers as described earlier (17). The following primers and probes were used to assess FTO gene expression in human subcutaneous fat tissue: tctgacccccaaagatgatg (forward), ctcggagaattagtttaggatatttca (reverse), and probe #59 (cat. no. 04688562001). Hipoxanthine phosphoribosyl transferase and -actin (ACTB) expression were measured as reference using HPRT primers tgaccttgatttattttgcatacc (forward), cgagcaagacgttcagtcct (reverse), and probe #73 (cat. no. 04688961001); and ACTB primers ccaaccgcgagaagatga (forward), ccagaggcgtacagggatag (reverse), and probe #64 (cat. no. 04688635001). PCR was performed for preincubation by 10 min at 95C, 45 cycles for 10 s at 95C, 30 s at 60C, and 30 s at 72C. All designs were assed for linearity and reproducibility. Relative quantification of gene expression was calculated with the LightCycler Software 4.0, using an arbitrary selected calibrator sample analyzed in all runs. Statistical analyses Because the number of risk-allele homozygotes was reduced, the additive and dominant models were considered the most appropriate. Covariance analysis was used to construct a model for quantitative traits where age, gender, and BMI (when appropriate) were included as covariates, and genotype was included as a fixed factor (GLM Univariate). Because fasting serum insulin/triglyceride levels and HOMA indices were not normally distributed, they were log transformed for analysis (SPSS, version 10.0, statistical package; SPSS, Chicago, IL). Pairwise linkage disequilibrium between SNPs was calculated by direct correlation (r2) using Haploview (version 3.2) (18). The AdmixMap program was used to test the possible effect of population stratification on associations with obesity and T2D (19,20). Because the Mexican-Mestizo population derived mainly from Amerindian and European (Spanish) populations, the model used included two primary parental populations. AdmixMap fits a logistic regression model of the trait on individual admixture, and it allows the inclusion of covariates such age, sex, and BMI. Previously reported Amerindian and European ancestral allele frequencies were used for the analysis (15,16). The study power to detect association with obesity was estimated using the QUANTO software (version 1.2; http://hydra.usc.edu/GxE/) and reached 80% statistical power under additive or dominant models.
2297

articles
genetics
Table 1 Association study of FTO rs9939609 polymorphism with obesity and T2D
Genotype n (%) n Stratified on BMI levela Nonobese Class I/II obese Class III obese Stratified on T2D Non-T2D T2D
b

TT

TA

AA

A-allele frequency (%)

Additive model OR (95% CI) Padd

Dominant model OR (95% CI) Pdom

424 216 148

297 (70.0) 137 (63.4) 72 (48.6)

114 (26.9) 67 (31.0) 60 (40.5)

13 (3.1) 12 (5.6) 16 (10.8)

16.7 21.1 31.1 1.38 (1.021.86) 2.42 (1.713.44) 0.034 0.0000004 1.42 (1.012.06) 2.80 (1.804.34) 0.046 0.000008

213 243

139 (65.3) 165 (67.9)

66 (30.9) 69 (28.4)

8 (3.8) 9 (3.7)

19.2 17.9 0.91 (0.651.27) 0.568 0.87 (0.591.30) 0.501

Data are n (%). P values and ORs calculated by logistic regression analysis using additive (add) and dominant (dom) models. CI, confidence interval; FTO, fat mass and obesity-associated gene; OR, odds ratio. a Adjusted for age, sex and institution where sample was collected. bAdjusted for age, sex, BMI, and institution where sample was collected. To reduce the probability of including control individuals who may develop diabetes type 2 (T2D) in later life, only unrelated nondiabetic subjects (fasting glucose levels <110 mg/dl) with no family history of diabetes aged 5080 years comprised the control group. Nonobese: BMI <30kg/m2; class I/II obese: 30 BMI <40kg/m2; class III obese >40kg/m2.

Table 2Clinical and biochemical parameters in obese nondiabetic individuals for the FTO gene rs9939609 polymorphism
TT N Males (%) Age (years) BMI (kg/m2)a Waist (cm)
a

TA 91 27.5 44.8 13.5 40.0 8.9 116.5 17.2 97.4 10.7 17.6 10.9 126.6 48.3 68.1 38.3 190.9 98.6 206.4 38.9 44.2 12.0

AA 23 53.3 41.5 7.6 43.1 9.9 124.0 24.1 94.7 10.7 20.1 15.4 143.3 59.8 64.7 36.9 194.5 111.1 221.9 53.7 44.6 10.6

Padd

Pdom

157 22.9 43.8 12.9 38.5 7.8 112.8 17.5 97.4 11.6 16.9 9.1 125.1 45.7 73.8 54.9 188.2 105.1 200.7 39.9 42.7 10.1
b

0.024 0.021 0.214 0.936 0.760 0.908 0.843 0.008 0.180

0.016 0.018 0.308 0.860 0.996 0.790 0.615 0.015 0.105

Fasting glucose (mg/dl)b Fasting insulin (U/ml) HOMA-B (%)b HOMA-S (%)
b

Triglicerides (mg/dl) Total cholesterol (mg/dl)b HDL-C (mg/dl)

b

Data are means s.d. unless otherwise indicated. All analyses were made using additive (add) and dominant (dom) models. Statistical power was 0.84 for the add model, and 0.73 for the dom model. FTO, fat mass and obesity-associated gene; HDL-C, high-density lipoprotein cholesterol; HOMA-B, homeostasis model assessment of B-cell function; HOMA-S, homeostasis model assessment of insulin sensitivity. a Adjusted for age and sex. bAdjusted for age, sex, and BMI.

Relative quantitative real-time PCR data are presented as means s.d. Comparisons of mRNA levels between lean vs. class I/II obese or class III obese subjects and with rs9939609 genotypes were performed using the KruskalWallis test. The threshold of significance was set at P < 0.05.
Results

Overall, 424 of the 788 Mexican mestizos were obese (148 with class III obesity), and 364 subjects were nonobese, whereas 243 subjects had T2D. Risk-allele frequencies for rs9939609, rs1421085, and rs17817449 SNPs were 20.5, 20.4, and 21.1%, respectively. Because linkage disequilibrium between these SNPs was high (r2 >0.85), only rs9939609 was considered for further analyses. Because risk genotype frequencies showed no significant differences in lean and overweight individuals (Padd = 0.368 and Pdom = 0.295), both groups were considered together as nonobese subjects for further analyses. Table1 shows the association with class III obesity was highly significant (P = 0.0000004
2298

and 0.000008 under the additive and dominant models, respectively), and these associations remained significant after adjusting for admixture (P= 0.000003 and 0.00009, respectively). The association was less significant for class I/II obesity (Padd=0.034 and Pdom = 0.046, and Padd = 0.021 and Pdom = 0.038 after adjusting for admixture). We also tested a possible association with T2D. Risk genotype frequencies showed no significant differences between T2D and nondiabetic individuals (Padd = 0.537 and Pdom = 0.658). As shown in Table 1, no association with T2D was found even when only nondiabetic individuals aged 5080 years were included as controls (Padd = 0.568 and Pdom = 0.501, and Padd = 0.531 and Pdom = 0.494 after adjusting for admixture). We also sequenced all 9 FTO exons and exonintron junctions from 15 class III obese individuals (nine bearing and six not bearing the risk alleles), but did not identify any FTO gene mutation.
VOLUME 16 NUMBER 10 | OCTOBER 2008 | www.obesityjournal.org

articles
genetics
Table 3Clinical and biochemical parameters in nonobese nondiabetic individuals for the FTO gene rs9939609 polymorphism
TT N Males (%) Age (years) BMI (kg/m2)a Waist (cm)
a

TA 81 46.9 51.7 12.4 24.3 2.3 83.6 9.2 91.4 13.4 6.3 3.9 72.8 28.2 175.6 73.2 155.2 86.8 208.4 40.2 50.2 14.7

AA 9 22.2 50.8 12.7 24.6 3.3 84.1 9.6 87.4 11.3 6.9 6.2 80.1 40.9 196.6 100.7 136.8 54.2 207.7 17.6 59.5 18.8

Padd

Pdom

181 32.6 52.2 13.4 24.5 2.4 85.9 9.6 89.7 11.3 7.8 5.7 87.1 45.5 158.8 92.2 160.6 91.9
b b

0.606 0.134 0.618 0.121 0.073 0.138 0.863 0.766 0.389

0.578 0.066 0.474 0.031 0.023 0.049 0.608 0.484 0.694

Fasting glucose (mg/dl)b Fasting insulin (U/ml) HOMA-B (%)b HOMA-S (%)
b

Triglycerides (mg/dl)b Total cholesterol (mg/dl) HDL-C (mg/dl)

b

214.8 44.5 52.5 12.8

Data are means s.d. unless otherwise indicated. All analyses were made using additive (add) and dominant (dom) models. Statistical power was 0.53 for the add model, and 0.43 for the dom model. FTO, fat mass and obesity-associated gene; HDL-C, high-density lipoprotein cholesterol; HOMA-B, homeostasis model assessment of B-cell function; HOMA-S, homeostasis model assessment of insulin sensitivity. a Adjusted for age and sex. bAdjusted for age, sex and BMI. 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 Lean Class I/II obese Class III obese

Figure 1 FTO mRNA expression levels in subcutaneous adipose tissue from lean (n = 10) vs. class I/II obese and class III obese individuals (n = 10 and 11, respectively). FTO gene expression readings were normalized to hipoxanthine phosphoribosyl transferase and -actin mRNA levels. Data presented are values relative to FTO mRNA levels observed in a single adipose tissue sample. Experiments were performed by duplicate (*P < 0.05 compared with lean subjects). Valuesare expressed as means s.d. FTO, fat mass and obesity-associated gene.

Table 2 compares anthropometric and biochemical measurements in obese non-T2D subjects according to genotype. Significantly higher BMI, waist circumference, and total cholesterol levels were observed in obese FTO rs9939609 A-allele carriers (Padd = 0.024, 0.021, and 0.008, respectively). Similar trends were observed in T2D patients (Padd <0.0001, <0.0001, and 0.024, for BMI, waist circumference, and total cholesterol levels, respectively). Although no differences in cholesterol levels were observed in nonobese non-T2D subjects, the risk allele showed a nominal association with lower insulin levels and HOMA-B, as well as higher HOMA-S in this subgroup (Pdom = 0.031, 0.023, and 0.049, respectively) (Table 3). We then genotyped the rs9939609, rs1421085, and rs17817449 polymorphisms in three Mexican-Amerindian populations (40 Yaqui, 24 Purpecha, and 48 Zapotec
obesity | VOLUME 16 NUMBER 10 | OCTOBER 2008

individuals). Although risk-allele frequencies varied among the three Amerindian groups (13.8, 12.5, and 1.0% for Yaqui, Purpecha, and Zapotec populations, respectively), risk alleles were always less frequent than in Mexican Mestizos. In spite of allele frequency differences, linkage disequilibrium between these SNPs was very similar to that found in Mexican Mestizos (r2 > 0.9). Furthermore, risk alleles were associated with higher BMI (31.9 7.2kg/m2 vs. 27.2 6.8kg/m2; P = 0.079) and waist circumference (105.9 15.8 cm vs. 93.1 12.6 cm; P = 0.017) under both models in Yaquis, which was the only Amerindian population including class III obese individuals. We analyzed relative FTO mRNA expression levels in human subcutaneous adipose tissue biopsies. Figure 1 shows that although FTO mRNA levels were higher in class I/II and class III obese individuals than in lean individuals, the difference was only significant for class III obesity (P = 0.043). Interestingly, rs9939609 TA genotype was significantly associated with higher FTO expression in the class III obesity group (mean for TA individuals (n = 4): 2.41 0.53 arbitrary units; mean for TT subjects (n = 3): 1.46 0.36 arbitrary units; P = 0.047).
Discussion

Relative mRNA FTO abundance (artbitrary units)

Obesity is reaching epidemic proportions and has become a serious health problem as it contributes to the increasing burden of T2D, heart disease, hypertension, and dyslipidemia(21). Although numerous gene variants have been associated with obesity or obesity-related phenotypes, very few have been consistently replicated (21,22). In agreement with several reports, mainly in white populations (28), we observed that FTO polymorphisms confer a highly significant association with obesity, particularly with the class III type, suggesting that FTO is a major susceptibility gene for obesity in the Mexican population. However, unlike previous observations that FTO predisposes to diabetes through an effect on BMI in several
2299

articles
genetics
very large cohorts (2), we found no association with T2D even without adjusting for BMI. In the present study, the proportion of obese individuals was similar in both diabetic and nondiabetic individuals (38.1 and 31.5%, respectively), and this may explain the lack of association. It could also be argued that the sample size is reduced; however, it should be pointed out that a highly significant association of a less frequent ABCA1 variant (11%) was significantly associated with T2D in the same sample (23). The prevalence of overweight and obesity in the Mexican population is one of the highest worldwide (70%), although the prevalence of class III obesity (2.5%) is lower than in whites(24). The FTO risk alleles frequencies in the overall Mexican population (20.421.1%) were lower than in whites (44.745%) and higher than in three Mexican-Amerindian populations (113.8%), as would be expected because Mexican-Mestizo population is the result of admixture mainly between native American and European (Spanish) populations. However, because the association with obesity remained significant after adjusting for admixture, it is unlikely that population stratification confounded our analysis. Associations of FTO variants with several metabolic parameters have been sought by several authors, with inconsistent results. In a large sample of white individuals, Freathy et al. (25) found an association of the rs9939609 A allele with increased insulin, glucose, and triglyceride levels, and with lower highdensity lipoprotein cholesterol levels. However, no evidence of these associations was observed when adjusting for BMI. In the present study, when analyzing obese and nonobese individuals separately, the FTO rs9939609 A-allele was associated with increased total cholesterol only in obese subjects (nondiabetic and diabetic) and with higher HOMA-S only in nonobese nondiabetic subjects. This allele was also associated with higher insulin sensitivity in a group of Chinese Han individuals with mean BMI = 24.2kg/m2, though the authors do not discuss this result (9). Together, these findings suggest that in the absence of obesity, the risk allele may be associated with increased insulin sensitivity, although this issue requires further study. Conversely, Andreasen et al. (5) observed that the effect of the FTO rs9939609 polymorphism on BMI levels is enhanced by low insulin sensitivity indices, particularly among homozygous A-allele carriers, suggesting that the impact of this polymorphism is highly influenced by insulin sensitivity. The possible interaction between FTO, obesity, insulin sensitivity, and other metabolic parameters needs to be defined in further studies. Although FTO is widely expressed in adult and fetal tissues(2), there is very scarce information on the function of this 2-oxoglutarate-dependent nucleic acid demethylase in different cell types. In agreement with the study of Whln et al. (26), we found that FTO expression in subcutaneous fat tissue was higher in obese than in lean subjects. However, in contrast with the lack of association with FTO expression according to genotype described by the latter authors (26), we found that FTO gene expression is higher in individuals bearing risk alleles only in the class III obesity group, suggesting that these SNPs are putatively functional or could be in linkage
2300

disequilibrium with other functional variants. Differences in the proportion of class III obese subjects in both studies may explain this difference. In conclusion, we found that FTO is a major risk factor for obesity (particularly class III) in the Mexican-Mestizo population, and is upregulated in subcutaneous adipose tissue of class III obese individuals.
Acknowledgments
This research was supported by grant no. 47414 from the Consejo Nacional de Ciencia y Tecnologa. We thank Salvador Ramrez-Jimenez, Maribel Rodriguez, and Luz E. Guilln-Pineda for their technical assistance.

Disclosure
The authors declared no conflict of interest.
2008 The Obesity Society

References
1. Must A, Spadano J, Coakley EH et al. The disease burden associated with overweight and obesity. JAMA 1999;282:15231529. 2. Frayling TM, Timpson NJ, Weedon MN et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 2007;316:889894. 3. Dina C, Meyre D, Gallina S et al. Variation in FTO contributes to childhood obesity and severe adult obesity. Nat Genet 2007;39:724726. 4. Saunders CL, Chiodini BD, Sham P et al. Meta-analysis of genome-wide linkage studies in BMI and obesity. Obesity (Silver Spring) 2007;15: 22632275. 5. Andreasen CH, Stender-Petersen KL, Mogensen MS et al. Low physical activity accentuates the effect of the FTO rs9939609 polymorphism on body fat accumulation. Diabetes 2008;57:95101. 6. Scuteri A, Sanna S, Chen WM et al. Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet 2007;3:e115. 7. Hinney A,nguyen TT, Scherag A et al. Genome wide association (GWA) study for early onset extreme obesity supports the role of fat mass and obesity associated gene (FTO) variants. PLoS ONE 2007;2:e1361. 8. Peeters A, Beckers S, Verrijken A et al. Variants in the FTO gene are associated with common obesity in the Belgian population. Mol Genet Metab 2008;93:481484. 9. Li H, Wu Y, Loos RJ et al. Variants in the fat mass- and obesity-associated (FTO) gene are not associated with obesity in a Chinese Han population. Diabetes 2008;57:264268. 10. Ohashi J, Naka I, Kimura R et al. FTO polymorphisms in oceanic populations. J Hum Genet 2007;52:10311035. 11. Gerken T, Girard CA, Tung YC et al. The obesity-associated FTO gene encodes a 2-oxoglutarate-dependent nucleic acid demethylase. Science 2007;318:14691472. 12. The Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 2003; 26:S520. 13. Villarreal-Molina MT, Aguilar-Salinas CA, Rodrguez-Cruz M et al. The ATP-binding cassette transporter A1 R230C variant affects HDL cholesterol levels and BMI in the Mexican population: association with obesity and obesity-related comorbidities. Diabetes 2007;56:18811887. 14. Levy JC, Matthews DR, Hermans MP. Correct homeostasis model assessment (HOMA) evaluation uses the computer program. Diabetes Care 1998;21:21912192. 15. Bonilla C, Parra EJ, Pfaff CL et al. Admixture in the Hispanics of the San Luis Valley, Colorado, and its implications for complex trait gene mapping. Ann Intern Med 2004;68:139153. 16. Choudhry S, Coyle NE, Tang H et al. Genetics of Asthma in Latino Americans (GALA) study: population stratification confounds genetic association studies among Latinos. Hum Genet 2006;118: 652664. 17. Mouritzen P, Noerholm M, Nielsen PS et al. ProbeLibrary: a new method for faster design and execution of quantitative real-time PCR. Nat Methods 2005;2:313316.
VOLUME 16 NUMBER 10 | OCTOBER 2008 | www.obesityjournal.org

articles
genetics
18. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformaticls 2005;21:263265. 19. Hoggart CJ, Parra EJ, Shriver MD et al. Control of confounding of genetic associations in stratified populations. Am J Hum Genet 2003;72: 14921504. 20. Hoggart CJ, Shriver MD, Kittles RA, Clayton DG, McKeigue PM. Design and analysis of admixture mapping studies. Am J Hum Genet 2004;74: 965978. 21. Bell CG, Walley AJ, Froguel P. The genetics of human obesity. Nat Rev Genet 2005;6:221234. 22. Rankinen T, Zuberi A, Chagnon YC et al. The human obesity gene map: the 2005 update. Obesity (Silver Spring) 2006;14:529644. 23. Villarreal-Molina MT, Flores-Dorantes MT, Arellano-Campos O et al. Association of the ATP-binding cassette transporter A1 R230C variant with early-onset type 2 diabetes in a Mexican population. Diabetes 2008;57:509513. 24. Secretara de Salud, Instituto Nacional de Salud Pblica. Encuesta Nacional de Salud y Nutricin 2005 (ENSANut), 2006. 25. Freathy RM, Timpson NJ, Lawlor DA et al. Common variation in the FTO gene alters diabetes-related metabolic traits to the extent expected given its effect on BMI. Diabetes 2008;57:14191426. 26. Whln K, Sjlin E, Hoffstedt J. The common rs9939609 gene variant of the fat mass and obesity associated gene (FTO) is related to fat cell lipolysis. J Lipid Res 2008;49:607611.

obesity | VOLUME 16 NUMBER 10 | OCTOBER 2008

2301