SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective

The peroxisome proliferator-activated receptor gamma 2 (PPARG2) gene has been intensively studied with relation to obesity and metabolic disorders. Indeed, a large number of studies assessing the association between the PPARG2 polymorphism Pro12Ala (rs1801282) and body mass index (BMI) have been published with some controversial results. In this meta-analysis, the effects of Pro12Ala polymorphism of the PPARG2 gene on BMI were investigated.

Design and Methods

Externally published data were collected and we included our own novel data from a study in the elderly participants (>55 years) of a Mediterranean cohort, the SUN (“Seguimiento Universidad de Navarra”) Project (n = 972). A total of 75 independent studies with 49,092 subjects (39,806 with the genotype Pro12Pro and 9,286 carrier subjects of the Ala allele) were included.

Results

The meta-analysis revealed a higher BMI with an overall estimation of +0.065 kg/m2 (95%CI = 0.026-0.103, P = 0.001) for homo-/heterozygous carriers of the Ala allele of the PPARG2 gene in comparison to non-carriers. The analysis also showed that there was heterogeneity (P for heterogeneity <0.001), but funnel plots did not suggest apparent publication bias. Furthermore, the association between the Pro12Ala polymorphism of the PPARG2 gene and increased BMI was stronger in Caucasian. Thus, carriers of the Ala allele had significantly higher BMI than non-carriers in a subsample of 6,528 Caucasian male subjects (standardized mean difference = 0.090, 95%CI=0.032-0.148, P = 0.002, P for heterogeneity = 0.121).

Conclusion

This updated meta-analysis showed that carriers of the Ala12 allele of the PPARG2 gene had a higher average BMI.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The peroxisome proliferator-activated receptor gamma 2 (PPARG2) is a nuclear receptor expressed mainly in adipose tissue that exerts an essential role in the regulation of adipocyte differentiation, lipid storage, and insulin sensitization [1]. PPARG2 also plays a key role in the entraining of adipose tissue lipid metabolism to nutritional state. The PPARG2 gene activation leads to upregulation of genes that mediate fatty acid uptake and trapping [2].

The PPARG2 gene is located in the chromosome 3p25; the Pro12Ala gene variant (rs1801282) of this gene, a missense mutation on exon B highly prevalent in the Caucasian population, has been controversially associated with obesity risk [3]. This mutation is a C[RIGHTWARDS ARROW]G substitution that results in the conversion of proline to alanine at residue 12 of the PPARG2 protein [3]. Functional analysis revealed that the receptor expressing this allele displays reduced deoxyribonucleic acid (DNA)-binding affinity and impaired transcriptional activity in target genes [4].

Two previous meta-analyses assessing the role of Pro12Ala of the PPARG2 gene on BMI and diabetes-related traits have been published [5]. In 2003 Masud et al. [5] conducted a meta-analysis to explore the effect of this Pro12Ala gene variant (rs1801282) on BMI in 19,136 subjects from 30 studies. They found a stronger association between the Ala12 allele and BMI in subjects with BMI≥27 kg/m2, whereas this association was not detected in individuals with BMI<27 kg/m2. In 2006, Tönjes et al. [6] performed another meta-analysis on the effect of the same SNP (rs1801282) of PPARG2 on diabetes-related traits in pre-diabetic subjects. They showed a direct association between carrying the Ala allele and greater BMI in 28,734 subjects from 45 studies.

A number of studies evaluating the association between the Pro12Ala polymorphism of the PPARG2 gene and BMI have been reported since 1998. In the present study, we pooled data obtained after a comprehensive and systematic literature review to overcome the limitations of single-study research work. Meta-analysis provides more reliable results by reducing the probability that random errors will produce false-positive or false-negative associations [7]. In our meta-analysis, we included novel data (n = 972) from a Mediterranean study (older participants of the SUN “Seguimiento Universidad de Navarra” cohort).

Subjects and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Subjects

This study was conducted within the framework of the SUN Project [8]. This is a dynamic cohort study, including only university graduates initiated in December 1999 in Spain. The survey and procedures have been previously described elsewhere [8]. For this study, elderly participants of the SUN project, those aged >55 years old when the basal questionnaire was completed, were invited to participate. Each volunteer received a kit designed to collect saliva (Oragene®ADN Self-Collector kit-OG250). Anthropometric data were collected from the baseline questionnaire. Self-reported information on BMI had been previously validated in a subsample of the SUN Project [10].

Voluntary completion of the first questionnaire was considered to imply informed consent to participate in the SUN Project, and written informed consent was requested to collaborate in this study. The study protocol was performed in accordance with the ethical standards of the Declaration of Helsinki (as revised in Hong Kong in 1989, in Edinburgh in 2000, and in South Korea in 2008) and was approved by the Institutional Ethical Review Board of the University of Navarra.

Genotyping

DNA was extracted from saliva samples according to the instructions of the kits manufacture (Oragene®ADN Self-Collector kit-OG250). The genotyping for the Pro12Ala polymorphism of the PPARG2 gene (rs1801282) was performed using a TaqMan assay with allele-specific probes on the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA) according to the standardized laboratory protocols. Replicate quality control samples were included in each genotyping plate with >99% of concordance.

Statistical analysis

A chi-square test was use to evaluate the Hardy-Weinberg equilibrium. The association of Pro12Ala polymorphism of the PPARG2 gene on BMI was analyzed with Student's t-test. The threshold for statistical significance was set a priori at P < 0.05.

A logistic regression model was performed to calculate Odds Ratios for obesity risk in the SUN cohort after adjustment for confounding factors, such as sex, age (years), physical activity practice (METs/h-week), and total energy intake (kcal/day).

Raw and adjusted geometric means for BMI have also been calculated to take into account skewed data. Linear general models were used to estimate adjusted BMI differences after controlling for the indicated confounding variables.

Meta-analysis

In our meta-analysis, the presence/absence of the Ala allele (dominant effect) was considered as the exposure, whereas differences in BMI between carriers and non-carriers of the Ala allele were considered as the outcome [7]. To systematically review differences in BMI across the presence/absence of the Pro12Ala polymorphism of the PPARG2 gene, we used a formal meta-analysis and updated the existing literature review [5]. A Pubmed search was done to find articles concerning the influence of Pro12Ala polymorphism of the PPARG2 gene on BMI using different combinations of the following search criteria: “BMI”, “obesity”, “Pro12Ala”, “rs1801282”. A total of 75 studies, comprising 109 samples and published before January 2011 were identified. Thirty of them were previously included in the meta-analisis by Masud (5), with a total sample of 19,136 subjects. The meta-analysis by Tönjes (6) provided 19,041 new subjects from 27 new studies, and finally, our work contributes with 11,243 novel individuals from 19 studies. They are mainly Caucasian, but also some Asian, Mexican-Hispanic and African-American subjects. Studies conducted in children or adolescents were not included as well as populations in which the allele prevalence of the Pro12Ala polymorphism of PPARG2 was not similar to those described in the HapMap database [11-14].

Table 1. Prevalence of the Pro12Ala polymorphism of the PPARG2 gene in elderly participants of the SUN cohort according to obesity status
 BMI ≥ 30 (kg/m2)BMI < 30 (kg/m2)ORaCI 95%P
  1. a

    OR for obesity risk adjusted by sex, age (years), physical activity practice (METs-h/week) and total energy intake (kcal/day)

Pro12Pro75 (0.76)739 (0.84)1 (ref)  
Pro12Ala22 (0.22)134 (0.15)1.600.955-2.6710.074
Ala12Ala2 (0.02)7 (0.01)3.010.602-15.0530.179
Ala carriers24 (0.24)141 (0.16)1.661.011-2.7380.045

No BMI differences were observed between carrier and non-carrier subjects of the Ala allele of the PPARG2 gene (Table 2).

Table 2. Characteristics of elderly participants in the SUN cohort according to the Pro12Ala polymorphism of the PPARG2 gene
 Pro12Pro (n = 814)Ala carriers (n = 164)P
  1. Data are shown as mean (95%CI). Continuous variables were compared using a Student's t test. Categorical variables were compared using chi-squared test.

  2. a

    General linear model after adjustment for sex, age (years), physical activity practice (METs-h/week), and total energy intake (kcal/day).

  3. b

    Geometric means for BMI.

Sex (% male)70730.329
Age (years)68.8 (68.4-69.2)70.0 (68.8-70.8)0.071
BMI (kg/m2)25.73 (25.51-25.95)26.19 (25.69-26.69)0.091
Adjusted BMI (kg/m2)a25.74 (25.53-25.92)26.13 (25.66-26.60)0.139
BMI (kg/m2)b25.54 (25.33-25.51)25.99 (25.52-26.48)0.089
Adjusted BMI (kg/m2)a, b25.56 (25.51-25.76)25.93 (25.48-26.39)0.141
Table 3. Meta-analyses of the Pro12Ala polymorphism of PPARG2 gene on BMI conducted in different population groups
 No. of casesNo. of controlsSMD95% CIPI2 (%)Pheterogeneity
  1. SMD: Pooled standardized mean differences for BMI (kg/m2) between 12Ala carriers and Pro12Pro subjects (dominant model).

  2. I2: percentage of the total variability in a set of effect sizes due to true heterogeneity.

  3. *: P value < 0.05 after correcting for Benjamini-Hochberg multiple comparisons.

All studies9,28639,8060.0650.026-0.1030.001054.1< 0.001
Sex       
 Women1,0344,1370.0730.003-0.1420.039414.10.275
 Men1,5445,3020.0980.041-0.1550.0008*34.50.054
Diabetes       
 Diabetic7663,9320.057-0.023-0.1360.161512.80.310
 Non diabetic3,82215,9060.053-0.003-0.1080.062944.30.001
Asian3084,1670.141-0.036-0.3170.118847.80.024
Caucasian7,97927,1470.0460.008-0.0840.016845.4< 0.001
Sex       
 Women9393,4970.0820.010-0.1550.026418.20.246
 Men1,5165,0120.0900.032-0.1480.0023*27.40.121
Diabetes       
 Diabetic6462,5590.058-0.029-0.1450.193631.00.161
 Non-diabetic3,41711,9560.043-0.012-0.0980.123638.80.010
BMI       
 Obese1,4445,1580.1560.041-0.2710.008168.8< 0.001
 Non-obese5,09316,3450.031-0.016-0.0770.197143.30.002
        
Table 4. Brief description of each population included in the meta-analysis
StudyMasud's meta- analysisTonjes' meta- analysisGalbete's meta-analysisN ControlsN Cases% Ala12Subject description
Beamer et al., 1998   1412816.57Obese, non diabetic Caucasian men and women
Beamer et al., 1998   40810921.08Non obese, non diabetic Caucasian men and women
Deeb et al., 1998   2577622.82Non diabetic Caucasian
Deeb et al., 1998   69527828.57Caucasian
Mori et al., 1998   203125.58Asian (Japanese)
Be et al., 1999   54021228.19Obese Caucasian men
Be et al., 1999   64122826.24Non obese Caucasian
Koch et al., 1999   753330.56Non obese, non diabetic Caucasian
Mancini et al., 1999   1141712.98Non obese, diabetic Caucasian men
Mancini et al., 1999   2555718.27Non obese, non diabetic, Caucasian
Ringel et al., 1999   38813425.67Non obese Caucasian
Ringel et al., 1999   37213126.04Non obese, diabetic Caucasian Subjects
Valve et al., 1999   1073424.11Obese, non diabetic, Caucasian women
Clement et al., 2000   2464916.61Diabetic Caucasian
Clement et al., 2000   2947820.97Non obese, non diabetic Caucasian
Clement et al., 2000   3396315.67Obese, non diabetic Caucasian
Cole et al., 2000   71121022.80Diabetic Caucasian
Hara et al., 2000   496458.32Asian (Japanese)
Hara et al., 2000   400153.61Asian (Japanese)
Hegele et al., 2000   902924.37Diabetic Oji-Cree (Canadian) women
Hegele et al., 2000   1482313.45Non diabetic Oji-Cree (Canadian) women
Lei et al., 2000   553437.21Asian (Taiwanese)
Meirhaeghe et al., 2000   66111121.12Non obese Caucasian
Meirhaeghe et al., 2000   1363420.00Obese Caucasian
Oh et al., 2000   211187.86Asian (Korean)
Poirier et al., 2000   50716824.89Non obese, non diabetic Caucasian men
Ek et al.,2001   45616025.97Non obese, non diabetic Caucasian men
Ek et al.,2001   2709425.82Non obese, non diabetic Caucasian men
Hseuh et al., 2001   2346622.00Mexican-American
Lindi et al., 2001   932621.85Non obese Caucasian
Luan et al., 2001   2035621.62Non obese, non diabetic Caucasian men
Luan et al., 2001   2656820.42Non obese, non diabetic Caucasian women
Nicklas et al., 2001   561420.00Obese, non diabetic Caucasian women
Schaffler et al., 2001   2768323.12Non obese Caucasian
Swarbrick et al., 2001   2157726.37Obese Caucasian
Swarbrick et al., 2001   2779425.34Non obese Caucasian
Ahluwalia et al., 2002   1394424.04Non obese, diabetic Caucasian
Doney et al., 2002   86923821.50Obese, diabetic Caucasian
Eriksson et al., 2002   32415231.93Non obese Caucasian
Frederiksen et al., 2002   1,67157425.57Non obese, non diabetic Caucasian
Gonzalez-Sanchez et al., 2002   371427.45Caucasian men
Gonzalez-Sanchez et al., 2002   821212.77Caucasian woman
Gonzalez-Sanchez et al., 2002   1372213.84Non obese Caucasian men
Gonzalez-Sanchez et al., 2002   1273119.62Non obese Caucasian women
Lindi et al., 2002   33715331.22Obese Caucasian
Masud et al., 2002   81327125.00Non obese Caucasian
Schneider et al., 2002   1563819.59Non obese, non diabetic Caucasian
Schneider et al., 2002   871313.00No obese, diabetic Caucasian
Stumvoll et al., 2002   1354223.73Non obese, non diabetic Caucasian
Stumvoll et al., 2002   39112824.66Non obese, non diabetic Caucasian
Thamer et al., 2002   732525.51Non obese Caucasian men
Yamamoto et al., 2002   454245.02Asian (Japanese)
Yamamoto et al., 2002   10986.84Asian (Japanese)
Yamamoto et al., 2002   7744.94Asian (Japanese)
Baratta et al., 2003   2964212.43Non diabetic Caucasian
Eurlings et al., 2003   572227.85Familiar combined hyperlipidemia Caucasian
Eurlings et al., 2003   933125.00Non diabetic Caucasian
Kahara et al., 2003   11764.88Asian (Japanese)
Kolehmainen et al., 2003   22826.67Obese Caucasian
Lindi et al., 2003   1143624.00Caucasian
Muller et al., 2003   67811714.72Pima Indian
Poulsen et al., 2003   1614722.60Caucasian
Poulsen et al., 2003   2687722.32Caucasian
Robitaille et al., 2003   58613418.61Caucasian
Rosmond et al., 2003   1868230.60Non obese Caucasian men
Thamer et al., 2003   50014822.84Caucasian
Andrulionyte et al., 2004   59217823.12Obese Caucasian
Buzzetti et al., 2004   1,00820717.04Obese, non diabetic Caucasian
Franks et al., 2004   862723.89Non diabetic Caucasian women
Franks et al., 2004   1142618.57Non diabetic Caucasian women
Franks et al., 2004   912219.47No diabetic Caucasian men
Franks et al., 2004   1083222.86No diabetic Caucasian men
Kim et al., 2004   977747.04Asian (Korean)
Pihlajamaki et al., 2004   20810333.12Non obese Caucasian
Pisabarro et al., 2004   39613.33No diabetic Caucasian
Tai et al., 2004   2,7962849.22Non diabetic Asian (Chinese, Malaya, Indian)
Tai et al., 2004   499397.25Impair glucose Asian (Chinese, Malaya, Indian)
Tai et al., 2004   3744610.95Diabetic Asian (Chinese, Malaya, Indian)
Takata et al., 2004   13974.79Non diabetic Asian men
Takata et al., 2004   901110.89Non diabetic Asian
Barbieri et al., 2005   3626715.62Non obese Caucasian
Danawati et al., 2005   19673.45Non diabetic Indian
Danawati et al., 2005   33072.08Diabetic Indian
Fornage et al., 2005   1,765794.28African-americans
Fornage et al., 2005   1,58147323.03Non obese Caucasian
Ghoussaini et al., 2005   67319222.20Non obese, non diabetic Caucasian
Ghoussaini et al., 2005   39711021.70Obese Caucasian
Meirhaeghe et al., 2005   89324021.18Caucasian
Mousavinasab et al., 2005   1737931.35Non diabetic Caucasian
Ostergard et al., 2005   611822.78Non diabetic, non obese Caucasian
Rhee et al., 2005   2262710.67Asian (Korea)
Tanko et al., 2005   1,08838626.19Non obese, non diabetic Caucasian women
Weiss et al., 2005   24825.00Non diabetic Caucasian men
Weiss et al., 2005   3749.76Non diabetic Caucasian women
Stefanski et al., 2006   1546028.04Obese, diabetic Caucasian
Canizales-Quinteros et al., 2007   1052619.85Non diabetic Amerindian-Mexican (BMI<25)
Canizales-Quinteros et al., 2007   763229.63Non diabetic Amerindian-Mexican (BMI>25)
Helwig et al., 2007   51519327.26Non obese, non diabetic Caucasian men
Kim et al., 2007   1151410.85Asian(Korea) women
Mattevi et al., 2007   1302315.03Non obese, non diabetic Caucasian men
Mattevi et al., 2007   1532915.93Non obese, no diabetic Caucasian women
Vaccaro et al., 2007   3014212.24Diabetic, obese Caucasian
Morini et al., 2008   5016511.48Non obese, non diabetic Caucasian men and women
Ben Ali et al., 2009   197188.37Tunisian women
Ben Ali et al., 2009   1512112.21Tunisian men
Ereqat et al.,2009   1792311.39Palestinian
Milewicz et al., 2009   2229630.19Non obese, non diabetic Caucasian women
Razquin et al., 2009   83713814.15Caucasian (high cardiovascular risk)
Present Study   81416516.85Caucasian

Our meta-analysis was carried out using the STATA 10.0 software (Stata-Corp, College Station, TX). We estimated a standardized mean difference (SMD) as the weighted effect size. This variable is the pooled estimate across studies for the BMI difference (kg/m2) in homo-/heterozygous carriers versus non-carriers of the Ala allele (dominant effect). We also calculated a 95% confidence interval (CI) for the pooled difference in BMI. A test of heterogeneity was also calculated, estimating the Cochran's Q statistic (Cochran 1954). A random-effect model was used when heterogeneity among studies was observed and a fixed-effect model when studies were homogenous. The I2 statistic was also used to evaluate heterogeneity among studies [15]. I2 describes the percentage of total variation across all the studies due to heterogeneity rather than chance and does not rely on the number of studies. Thus, it can be used for comparisons across meta-analyses with different number of studies. Percentages around 25% (I2=25), 50% (I2=50), and 75% (I2=75) would indicate low, medium, and high heterogeneity, respectively [15].

The meta-analysis was carried out first in the total sample, 9,286 carrier subjects of the Ala allele compared with 39,806 non-carrier subjects (dominant model) and because significant heterogeneity was evident, the DerSimonian and Laird's random effect model was used. To evaluate potential sources of heterogeneity, we separated samples into sub-groups using the following criteria: sex, ethnic group (Asian or Caucasian), and presence or absence of type 2 diabetes. Within the Caucasian samples, we studied separately men and women, lean and obese, and diabetic and non-diabetic subjects. When homogeneity between samples was observed, the fixed-effects model was used. The false discovery rate method from Benjamini and Hochberg was used to control for multiple testing in the sub-groups analyses [16].

The visual Funnel plots and the Egger's test were used to detect evidence of possible bias, resulting from selective publication of positive studies [18].

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Effect of the Pro12Ala polymorphism on the body mass index in an elderly SUN population

The frequency of the Ala allele of the PPARG2 gene was 0.19 in our elderly SUN population. Specifically, 83% of the subjects carried the Pro12Pro genotype (wild type), 16% of the subjects were heterozygous for the mutation (Pro12Ala), and only 1% was homozygous (Table 1). The allele distribution fulfilled the Hardy-Weinberg equilibrium. The presence of the Ala allele of the PPARG2 gene significantly increased obesity risk (OR=1.664, 95%CI=1.011-2.738, P = 0.045).

Meta-analysis

We pooled 109 comparisons from 75 independent studies, comprising 49,092 subjects: 39,806 subjects had the genotype Pro12Pro and 9,286 were carrier subjects of the Ala allele (Tables 3 and 4).

image

Figure 1. Meta-analysis of the Pro12Ala polymorphism of the PPARG2 gene effect on BMI comprising 49,092 subjects. aStudies included in Masud's meta-analysis. bStudies included in Tonjes' meta-analysis

Download figure to PowerPoint

The meta-analysis revealed a higher BMI with an overall estimation of +0.065 kg/m2 (95% CI = 0.026-0.103, P = 0.001, Figure 1) for carriers of the Ala allele of the PPARG2 gene in comparison to non-carriers (dominant effect). The analysis also showed that there was significant heterogeneity (P for heterogeneity < 0.001) among the different studies. The Funnel plot for all samples and the Egger's test (P = 0.249) showed symmetrical distribution, indicating that there was no apparent publication bias (Figure 2).

image

Figure 2. Funnel plot of 109 samples included in the meta-analysis, Egger's test: P = 0.249. SMD: standardized mean difference.

Download figure to PowerPoint

We investigated possible sources of heterogeneity such as sex, ethnic group (Asian or Caucasian population), and type 2 diabetes (presence or absence). Within the Caucasian samples, we studied separately men and women, lean and obese, and also diabetic and non-diabetic subjects.

As it has been reported the Pro12Ala polymorphism of the PPARG2 gene is linked to lower type 2 diabetes risk. We performed a meta-analysis, including 4,698 diabetic patients, but no effect for the Ala allele on BMI was apparent.

A significant association between Pro12Ala polymorphism of the PPARG2 gene and increased BMI was detected in studies performed separately in Caucasian men, Caucasian women, and in obese Caucasian subjects. Following the hypothesis that the Pro12Ala polymorphism of the PPARG2 gene may have a stronger effect on BMI in markedly obese individuals, we restricted the meta-analysis to Caucasian individuals and conducted two separate assessments: in obese (6,602 subjects with BMI ≥ 30 kg/m2) and non-obese subjects (21,438 subjects with BMI < 30 kg/m2). A significant association between the Pro12Ala polymorphism of the PPARG2 gene and BMI in obese subjects was observed (SMD = 0.156, 95%CI = 0.041-0.271, P = 0.008, P for heterogeneity < 0.001). Moreover, in a fixed-effects model meta-analysis with 6,528 Caucasian men, carriers of the Ala allele had significantly higher BMI than non-carriers (SMD = 0.090, 95%CI = 0.032-0.148, P = 0.002, P for heterogeneity = 0.121). Similar results were observed in Caucasian women (SMD = 0.082, 95%CI = 0.010-0.155, P = 0.026, P for heterogeneity = 0.246) although after the Benjamini-Hochberg multiple comparison the results did not remain statistically significant.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

PPARG2 is one of the most studied genes as potentially linked to obesity phenotypes. Indeed, a large number of human studies have shown that the Ala12 allele was associated with increased adiposity [4].

In our meta-analysis, we compiled previously reported studies and have also included novel data from a study in 972 older participants of the SUN Project. To reduce heterogeneity, different analyses were conducted for Caucasian or Asian population, men and women, diabetic and non-diabetic subjects, and also obese and non-obese subjects.

Our meta-analysis with a total of 49,092 subjects (39,806 Pro12Pro subjects and 9,286 subjects) revealed a significantly higher BMI with an overall estimation of +0.065 kg/m2 for homo-/heterozygous carriers of the Ala allele of the PPARG2 gene when compared to non-carriers (Pro12Pro subjects).

A similar effect of the Pro12Ala gene variant of the PPARG2 gene on BMI changes (+0.066 kg/m2) was reported in a former meta-analysis [5]. They found heterogeneity but no apparent publication bias.

The magnitude of BMI change observed in our study is more consistent within Caucasian men. However, it is quite modest when compared with the average BMI increment per risk allele (+0.170 kg/m2 for 32 genetic variants with P-values < 5×10−8) in the largest meta-analysis (249,796 individuals) of GWAS for BMI thus far published [19, 20]. Moreover, the following information for the Pro12Ala SNP (rs1801282) was obtained from the GIANT consortium data files: a minor allele frequency equal to 0.075 and a P value equal to 0.0193 (after using regression coefficients and correction for inflation) in 123,856 subjects. Unfortunately, specific information for BMI according to the presence of the Ala12 allele is not available in the GIANT study.

Sex-differences in studies on obesity and genetic variants are extensively described in the literature. It is known that PPARG2 expression levels are higher in subcutaneous than visceral adipose tissue [21]. And also it is worthy to mention that PPARG2 inhibits a key enzyme in estrogen biosynthesis the aromatase gene, suggesting that the action of PPARG2 could be modulated by sex steroids [22]. This fact may be a possible explanation for the differences observed between men and women [23]. Regardless of the biological pathways, it is important to consider that sex-specific lifestyle components (dietary intake, alcohol consumption, physical activity levels, or smoking habits) could act as modifiers for the effect of a given genetic variant on BMI.

Other studies have suggested that Pro12Ala polymorphism of the PPARG2 gene may exert its effect on BMI only in markedly obese individuals, for instance, those subjects with a BMI higher than 27 kg/m2 of the Masud's meta-analysis [5]. Our study partially confirmed this point since statistical differences were only found for obese Caucasian (BMI ≥ 30 kg/m2) subjects in a random model meta-analysis, although the potential effect of publication bias should also to be taken into consideration.

Our meta-analysis has strengths and limitations. The lack of information on environmental factors and also gene-gene interactions is a limitation of our work. It has been described that dietary modification may influence the association between genetic variants and obesity, and they should be accounted for [4].

One advantage is that information for more 49,000 subjects is pooled together. It derives from 75 independent studies, including 109 samples of different size from 30 to 3,080 participants, which may account as a source of heterogeneity. We included populations with Ala12 frequency similar to those described in the HapMap database [24]. The frequency of the Ala12 allele of the PPARG2 gene varies from 2% to 33% in our meta-analysis.

In some sub-groups analysis (i.e. after distributing subjects for the ethnic group or obesity status), heterogeneity remained, which suggests that there may be more than one source of heterogeneity at play. Heterogeneity usually accompanies genetic association studies. The level of heterogeneity in our meta-analyses by using the I2 test ranges from 13% to 69%, which corresponds to medium degree of heterogeneity [15].

Results from both random- and fixed-effects models have been provided throughout the article and did not differ substantially. The drawback of combining studies in the presence of heterogeneity is less related with the pooled estimate achieved and more with the explicit demonstration of this heterogeneity and the need for caution in interpretation.

Finally, to address the clinical relevance of the Ala allele of the PPARG2 gene information from weight-loss intervention studies is compiled here. Lindi et al. [25] and Franks et al. [26] showed that subjects with the Ala12 allele lost more weight after 3 years and 1-year follow-up period, respectively. But Adamo et al. [27] showed the opposite result: the Ala12 allele was more frequent in diet-resistant individuals. No effect was reported for Nicklas et al. [28] and Matsuo et al. [29]. They attributed the lack of effect to specific characteristic of the population: only women and a low prevalence of the Ala12 allele, respectively. Other relevant point refers to the effect of genomic information as a behavioral health intervention. A recent review observed that simple communication of genetic information and disease susceptibility, in some motivated groups, might be sufficient to trigger lifestyle changes; for other groups, additional strategies may be required [30]. Unfortunately, the effect of this genotype of PPARG2 on body weight regulation remains unclear.

In conclusion, the current meta-analysis showed that the Pro12Ala polymorphism of the PPARG2 gene has a modest role in increasing BMI, which cannot be considered very relevant from the clinical point of view; however, this positive association stronger among Caucasian men.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The scholarship to C. Galbete from the Asociación de Amigos de la Universidad de Navarra is fully acknowledged. E. Toledo was supported by a Rio Hortega post-residency fellowship of the Instituto de Salud Carlos III, Ministry of Economy and Competitiveness, Spanish Government.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Razquin C, Martinez JA, Martinez-Gonzalez MA, et al. The Mediterranean diet protects against waist circumference enlargement in 12Ala carriers for the PPARgamma gene: 2 years' follow-up of 774 subjects at high cardiovascular risk. Br J Nutr 2009;102:672-679.
  • 2
    Ereqat S, Nasereddin A, Azmi K, Abdeen Z, Amin R. Impact of the Pro12Ala polymorphism of the PPAR-gamma 2 gene on metabolic and clinical characteristics in the Palestinian type 2 diabetic patients. PPAR Res 2009;2009:874126.
  • 3
    Morini E, Tassi V, Capponi D, et al. Interaction between PPARgamma2 variants and gender on the modulation of body weight. Obesity (Silver Spring) 2008;16:1467-1470.
  • 4
    Razquin C, Marti A, Martinez JA. Evidences on three relevant obesogenes: MC4R, FTO and PPARgamma. Approaches for personalized nutrition. Mol Nutr Food Res 2011;55:136-149.
  • 5
    Masud S, Ye S. Effect of the peroxisome proliferator activated receptor-gamma gene Pro12Ala variant on body mass index: a meta-analysis. J Med Genet 2003;40:773-780.
  • 6
    Tonjes A, Scholz M, Loeffler M, Stumvoll M. Association of Pro12Ala polymorphism in peroxisome proliferator-activated receptor gamma with Pre-diabetic phenotypes: meta-analysis of 57 studies on nondiabetic individuals. Diabetes Care 2006;29:2489-2497.
  • 7
    Rothman K, Sander G, Timothy L. Meta-Analysis. Modern Epidemiology. (3rd ed.) Philadelphia: Lippincott Williams & Wilkins; 2008. p. 757.
  • 8
    Martinez-Gonzalez MA, Sanchez-Villegas A, De Irala J, Marti A, Martinez JA. Mediterranean diet and stroke: objectives and design of the SUN project. Seguimiento Universidad de Navarra. Nutr Neurosci 2002;5:65-73.
  • 9
    Segui-Gomez M, de la Fuente C, Vazquez Z, de Irala J, Martinez-Gonzalez MA. Cohort profile: the 'Seguimiento Universidad de Navarra' (SUN) study. Int J Epidemiol 2006;35:1417-1422.
  • 10
    Bes-Rastrollo M, Perez Valdivieso JR, Sánchez-Villegas A, Alonso A, Martinez-Gonzalez MA. Validación del peso e indice de masa corporal auto-declarados de los participantes de una cohorte de graduados universitarios. Rev Esp Obes 2005;183-189.
  • 11
    Chuang LM, Hsiung CA, Chen YD et al. Sibling-based association study of the PPARgamma2 Pro12Ala polymorphism and metabolic variables in Chinese and Japanese hypertension families: a SAPPHIRe study. Stanford Asian-Pacific Program in Hypertension and Insulin Resistance. J Mol Med (Berl) 2001;79:656-664.
  • 12
    Kuliczkowska J, Filus A, Trzmiel A et al. PPAR-gamma2 Pro12Ala polymorphism in the population of obese and non-obese men of the city of Wroclaw. Endokrynol Pol 2008;59:312-315.
  • 13
    Ostergard T, Ek J, Hamid Y et al. Influence of the PPAR-gamma2 Pro12Ala and ACE I/D polymorphisms on insulin sensitivity and training effects in healthy offspring of type 2 diabetic subjects. Horm Metab Res 2005;37:99-105.
  • 14
    Pisabarro RE, Sanguinetti C, Stoll M, Prendez D. High incidence of type 2 diabetes in peroxisome proliferator-activated receptor gamma2 Pro12Ala carriers exposed to a high chronic intake of trans fatty acids and saturated fatty acids. Diabetes Care 2004;27:2251-2252.
  • 15
    Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557-560.
  • 16
    Benjamini Y, Hochberg Y. Controlling the false discovert rate—a practical and powerful approach to multiple testing. J R Stat Soc Ser B-Methodol 1995;57:289-300.
  • 17
    Razquin C, Martinez JA, Martinez-Gonzalez MA et al. A Mediterranean diet rich in virgin olive oil may reverse the effects of the -174G/C IL6 gene variant on 3-year body weight change. Mol Nutr Food Res 2010;54 Suppl 1:S75-S82.
  • 18
    Golder S, Loke YK, Bland M. Meta-analyses of adverse effects data derived from randomised controlled trials as compared to observational studies: Methodological overview. PLoS Med 2011;8:e1001026.
  • 19
    Speakman JR, Levitsky DA, Allison DB et al. Set points, settling points and some alternative models: theoretical options to understand how genes and environments combine to regulate body adiposity. Dis Model Mech 2011;4:733-745.
  • 20
    Speliotes EK, Willer CJ, Berndt SI et al. Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nat Genet 2010;42:937-948.
  • 21
    Gonzalez Sanchez JL, Serrano Rios M, Fernandez Perez C, Laakso M, Martinez Larrad MT. Effect of the Pro12Ala polymorphism of the peroxisome proliferator-activated receptor gamma-2 gene on adiposity, insulin sensitivity and lipid profile in the Spanish population. Eur J Endocrinol 2002;147:495-501.
  • 22
    Memisoglu A, Hankinson SE, Manson JE, Colditz GA, Hunter DJ. Lack of association of the codon 12 polymorphism of the peroxisome proliferator-activated receptor gamma gene with breast cancer and body mass. Pharmacogenetics 2002;12:597-603.
  • 23
    Ben Ali S, Ben Yahia F, Sediri Y et al. Gender-specific effect of Pro12Ala polymorphism in peroxisome proliferator-activated receptor gamma-2 gene on obesity risk and leptin levels in a Tunisian population. Clin Biochem 2009;42:1642-1647.
  • 24
    Interantional HapMap Project. Available from: http://hapmap.ncbi.nlm.nih.gov/index.html.en.
  • 25
    Lindi VI, Uusitupa MI, Lindstrom J et al. Association of the Pro12Ala polymorphism in the PPAR-gamma2 gene with 3-year incidence of type 2 diabetes and body weight change in the Finnish Diabetes Prevention Study. Diabetes 2002;51:2581-2586.
  • 26
    Franks PW, Jablonski KA, Delahanty L et al. The Pro12Ala variant at the peroxisome proliferator-activated receptor gamma gene and change in obesity-related traits in the Diabetes Prevention Program. Diabetologia 2007;50:2451-2460.
  • 27
    Adamo KB, Dent R, Langefeld CD et al. Peroxisome proliferator-activated receptor gamma 2 and acyl-CoA synthetase 5 polymorphisms influence diet response. Obesity (Silver Spring) 2007;15:1068-1075.
  • 28
    Nicklas BJ, van Rossum EF, Berman DM et al. Genetic variation in the peroxisome proliferator-activated receptor-gamma2 gene (Pro12Ala) affects metabolic responses to weight loss and subsequent weight regain. Diabetes 2001;50:2172-2176.
  • 29
    Matsuo T, Nakata Y, Katayama Y et al. PPARG genotype accounts for part of individual variation in body weight reduction in response to calorie restriction. Obesity (Silver Spring) 2009;17:1924-1931.
  • 30
    Bloss CS, Madlensky L, Schork NJ, Topol EJ. Genomic information as a behavioral health intervention: Can it work? Per Med 2011;8:659-667.