Effects of established BMI-associated loci on obesity-related traits in a French representative population sample

Background Genome-wide association studies have identified variants associated with obesity-related traits, such as the body mass index (BMI). We sought to determine how the combination of 31 validated, BMI-associated loci contributes to obesity- and diabetes-related traits in a French population sample. The MONA LISA Lille study (1578 participants, aged 35–74) constitutes a representative sample of the population living in Lille (northern France). Genetic variants were considered both individually and combined into a genetic predisposition score (GPS). Results Individually, 25 of 31 SNPs showed directionally consistent effects on BMI. Four loci (FTO, FANCL, MTIF3 and NUDT3) reached nominal significance (p ≤ 0.05) for their association with anthropometric traits. When considering the combined effect of the 31 SNPs, each additional risk allele of the GPS was significantly associated with an increment in the mean [95% CI] BMI of 0.13 [0.07-0.20] kg/m2 (p = 6.3x10-5) and a 3% increase in the risk of obesity (p = 0.047). The GPS explained 1% of the variance in the BMI. Furthermore, the GPS was associated with higher fasting glycaemia (p = 0.04), insulinaemia (p = 0.008), HbA1c levels (p = 0.01) and HOMA-IR scores (p = 0.0003) and a greater risk of type 2 diabetes (OR [95% CI] = 1.06 [1.00-1.11], p = 0.03). However, these associations were no longer statistically significant after adjustment for BMI. Conclusion Our results show that the GPS was associated with a higher BMI and an insulin-resistant state (mediated by BMI) in a population in northern France.


Background
According to the World Health Organization (WHO)'s criterion for obesity (body mass index (BMI) ≥ 30 kg/m 2 ), up to 15% of the adults in Europe are obese [1]. The prevalence of obesity has more or less doubled since 1980 [2]. Obesity is a serious public health issue worldwide. Indeed, there is a well-documented relationship between a high BMI on one hand and mortality and morbidity due to chronic diseases (such as cardiovascular disease, certain cancers, type 2 diabetes (T2D) and osteoarthritis) on the other [3]. Accordingly, the WHO has declared obesity to be a global epidemic that affects both industrialized and non-industrialized countries [4].
Body fat mass is influenced by the combination of genetic factors and lifestyle factors (such as diet and physical activity). Family and twin studies have shown that genetic factors account for 40-70% of the population variation in BMI [5,6]; this may explain why people are not all equally affected by obesity in an obesogenic environment [7].
Genome-wide association studies (GWASs) have sought to elucidate the genetic basis of obesity and its related traits. To date, 32 genetic loci have been unequivocally associated with BMI [8]. Several studies have replicated these associations and have taken account of the combined impact of these GWAS-validated loci when considering BMI and other obesity-related phenotypes [8,9]. The objective of the present study was to replicate the combined effects of the established BMI-associated loci on BMI, body fat percentage, waist circumference, waist-to-hip ratio (WHR) and obesity risk in a representative sample of the general population in northern France (n = 1578). Furthermore, the high observed burden of obesity-related co-morbidities (such as insulin resistance and T2D) prompted us to test the impact of the BMI-associated loci on glucose-related traits and the risk of T2D.
Similar results were obtained after taking into account missing genotypes (Additional file 1: Table S4). Associations between the GPS and the waist and hip circumferences disappeared after further adjustment for BMI.
We also investigated the possible effect of interactions between the GPS and gender, physical activity (PA), smoking status and alcohol consumption on anthropometric variables but did not detect any significant interactions (data not shown).
To distinguish between the effects of the GPS and the effects of the covariables classically associated with BMI (age, gender, PA, smoking status and alcohol consumption), we compared the crude and adjusted models ( Table 2). The GPS alone accounted for 1% of the variance in the BMI, whereas the covariables accounted for 6%. Overall, the GPS and the covariables explained 7% of the variance in the BMI.
We also investigated the association between the GPS and the obesity risk. Each additional BMI-raising allele was associated with a 3% increase in the obesity risk (OR [95% CI] = 1.03 [1.00-1.07]; p = 0.047).
The genetic predisposition score, glucose-related traits and the type 2 diabetes risk Given that obesity is an important determinant of glycaemic traits and insulin resistance, we assessed the association between the GPS on one hand and fasting plasma glucose, HbA1c and insulin levels, the HOMA-IR and HOMA-B scores and the risk of T2D on the other. We detected significant associations between the GPS and higher fasting plasma glucose (β ± SE = +0.017 ± 0.008 mmol/L, p = 0.04), insulin (β ± SE = +0.14 ± 0.06 μIU/mL, p = 0.008) and HbA1c levels (β ± SE = +0.012 ± 0.005%, p = 0.01) and a higher HOMA-IR (β ± SE = +0.06 ± 0.02, p = 0.0003) ( Table 3). The GPS was also significantly associated with a higher risk of T2D (adjusted OR [95% CI] = 1.06 [1.00-1.11], p = 0.03). However, these associations were no longer statistically significant after adjustment for BMI.

Discussion
Although the MONA LISA Lille study's statistical power was too low (68%) to detect significant individual associations, 25 of the 31 investigated SNPs presented effects with the expected direction. Moreover, the effect alleles for the FTO rs9939609 and FANCL rs887912 SNPs were nominally associated with higher BMI. The GPS (corresponding to the cumulative contribution of the 31 validated BMI-associated SNPs) showed a significant, positive association with BMI. Each additional effect allele was associated with a mean increment of 0.13 kg/m 2 in the BMI (which corresponds to a weight increment of 376 g for a person measuring 1.70 m in height) and a 3% increase in the risk of obesity. The GPS was also significantly associated with body fat percentage and waist and hip circumferences, although the last two associations did not resist adjustment for BMI (suggesting that they were driven by overall general adiposity). The genetic susceptibility associated with the GPS explained only 1% of the variance in the BMI, whereas the combined effect of known lifestyle factors accounted for 6%. Although it is clear that (i) genetic factors account for 40-70% of the population variation in BMI and (ii) the 31 SNPs studied here have been robustly validated as BMI-susceptible variants in GWASs and replication studies, the SNPs' combined effect on BMI and the obesity risk was quite small. However, our results are in agreement with previous reports [8,10,11]. Gene-environment interactions may also account for variance in the BMI. Several studies have reported that PA is associated with a reduction in the GPS's impact on BMI [12,13]. Like others [12], we failed to detect significant interactions between the GPS and PA when considering several anthropometric traits (BMI, body fat percentage, waist and hip circumferences and WHR). Our failure to detect this interaction is probably due to the relatively small sample size. In fact, very large sample sizes are needed when exploring this type of interaction. For example, Ahmad et al. showed that a population size of 20,000 is required to detect a β GE interaction effect of −0.07 kg/m 2 [13].
Given that obesity is a major risk factor for insulin resistance [14], the accumulation of obesity risk alleles may alter glucose metabolism and predispose the individual to T2D. To evaluate this hypothesis, we looked at whether the GPS was associated with glucose-related variables and the T2D risk in the MONA LISA Lille study. Indeed, we found significant associations between the GPS on one hand and higher fasting plasma glucose, insulin and HbA1c levels and insulin resistance on the other. We also showed that each additional BMI-raising allele was associated with a 6% increment in the T2D risk. Our results in a general population sample are consistent with previous reports. In a French case-control study, each additional allele in the GPS was associated with higher insulin resistance and a 3% increase in the T2D risk [15]. In the EPIC prospective cohort study, each additional allele in the GPS was also associated with a 4% increase in the T2D risk [10]. In both these previous studies (as in the present study), all the statistically significant associations were abolished after adjustment for BMI -meaning that overall general adiposity explained the association between the GPS and insulin resistance or T2D.

Conclusions
Our results showed that the combination of common genetic variants was moderately associated with BMI and BMI-related variables in a sample of the general population from northern France. Despite the fact that the heritability of BMI is estimated to be 40-70% [5], the combination of 31 validated, BMI-associated loci only explained only 1% of the variance in the BMI (i.e. less than 2-4% of the heritability) [8]. Hence, characterization of this unexplained heritability requires other approaches.

Methods
The MONA LISA Lille study The MONA LISA (Monitoring National du Risque Artériel; National Monitoring of Arterial Risk) Lille study was a population-based, cross-sectional study of a representative sample of 1578 participants recruited from within the Lille urban area in northern France. In accordance with the French legislation on biomedical research, the study protocol was approved by the appropriate independent ethics committee (Comité Consultatif de Protection des Personnes dans la Recherche Biomédicale de Lille) and written informed consent was obtained from all participants. The study design and methods are described in the Additional file 1: methods. Briefly, anthropometric traits were recorded during a physical examination of each individual and a blood sample was collected (for DNA extraction and clinical biochemistry assays). The BMI was calculated according to the Quetelet equation. Obesity was defined as a BMI of 30 kg/m 2 or more. Type 2 diabetes was defined according to the 1997 American Diabetes Association definition (fasting plasma glucose ≥ 7.0 mmol/l and/or treatment for diabetes, including diet and/or oral antidiabetic drugs and/or insulin) [16].

Statistical analysis
Statistical analyses were performed with SAS 9.1 software (SAS Institute Inc., Cary, NC, USA). The Hardy-Weinberg equilibrium was tested using a χ 2 test with one degree of freedom.
The GPS was derived as described previously [17]. Briefly, a weighting method was used to calculate the GPS on the basis of 31 SNPs. Each SNP was weighted according to its relative effect size (i.e. the β coefficient). In order to measure the effect of each SNP on BMI with greater accuracy and precision, β coefficients were derived as described by Speliotes et al. [8]. We rescaled the weighted scores to reflect the number of risk alleles.
Hence, each point on the GPS corresponded to one risk allele. When calculating the GPS, missing genotype data were replaced with the average allele count for the corresponding SNPs. However, individuals with missing genotypes for more than 10% of the loci were excluded from the GPS analyses (n = 30).
We used general linear regression models to test the associations of individual BMI-related SNPs and the GPS with adiposity-related traits (including BMI, body fat percentage, WHR, waist circumference and hip circumference) and glucose-related traits (assuming an additive effect of the BMI-increasing alleles). A logistic regression model was used to test the association between the GPS and the risk of obesity or T2D. Interactions between the GPS on one hand and gender, PA, smoking status and alcohol consumption on the other were tested by including the GPS, interaction variables and the interaction terms (GPS x interaction variables) in general linear regression models.
The associations between genetic variants and BMI, obesity and interactions were adjusted for age, gender, smoking status, PA and alcohol consumption. The associations between genetic variants and body fat percentage, WHR, waist circumference and hip circumference were adjusted for age, gender, smoking status, PA, and alcohol consumption including or not BMI, depending of models. The associations between genetic variants and biological parameters and the T2D risk were adjusted for age, gender, BMI, smoking status, PA and alcohol consumption. Data distributions for plasma glucose and insulin levels and HOMA-IR and HOMA-B scores were normalized by log transformation.
Bonferroni correction was used to adjust for the Hardy-Weinberg equilibrium and for the multiple testing in the individual obesity-related trait analyses. The threshold for statistical significance was set to p ≤ 0.0016 (for 31 independent SNPs). Nominal significance was defined as 0.0016 < p < 0.05.
For the GPS analyses, the threshold for statistical significance was set to p ≤ 0.05.
The power calculations for association analyses (performed a priori using Quanto v1.2.4 software (http:// biostats.usc.edu/Quanto.html) on the basis of the mean BMI values from the MONA LISA Lille study and the effect allele frequencies and effect sizes originally reported by Speliotes et al. [8]) indicated that the statistical power of our study (for detecting a significant association between an individual SNP and BMI with a one-sided p value of 0.05) was 68%.
The power calculations for the GPS analysis were performed using the pwr package developed by Stéphane Champely. The statistical power for detecting significant association between GPS and BMI (using a p value at 0.05) was 99%.