Epigenetic associations of type 2 diabetes and BMI in an Arab population

Abstract

Background: The prevalence of type 2 diabetes (T2D) and obesity has dramatically increased within a few generations, reaching epidemic levels. In addition to genetic risk factors, epigenetic mechanisms triggered by changing environment are investigated for their role in the pathogenesis of these complex diseases. Epigenome-wide association studies (EWASs) have revealed significant associations of T2D, obesity, and BMI with DNA methylation. However, populations from the Middle East, where T2D and obesity rates are highest worldwide, have not been investigated so far.

Methods: We performed the first EWAS in an Arab population with T2D and BMI and attempted to replicate 47 EWAS associations previously reported in Caucasians. We used the Illumina Infinium HumanMethylation450 BeadChip to quantify DNA methylation in whole blood DNA from 123 subjects of 15 multigenerational families from Qatar. To investigate the effect of differing genetic background and environment on the epigenetic associations, we further assessed the effect of replicated loci in 810 twins from UK.

Results: Our EWAS suggested a novel association between T2D and cg06721411 (DQX1; p value = 1.18 × 10(-9)). We replicated in the Qatari population seven CpG associations with BMI (SOCS3, p value = 3.99 × 10(-6); SREBF1, p value = 4.33 × 10(-5); SBNO2, p value = 5.87 × 10(-5); CPT1A, p value = 7.99 × 10(-5); PRR5L, p value = 1.85 × 10(-4); cg03078551, intergenic region on chromosome 17; p value = 1.00 × 10(-3); LY6G6E, p value = 1.10 × 10(-3)) and one with T2D (TXNIP, p value = 2.46 × 10(-5)). All the associations were further confirmed in the UK cohort for both BMI and T2D. Meta-analysis increased the significance of the observed associations and revealed strong heterogeneity of the effect sizes (apart from CPT1A), although associations at these loci showed concordant direction in the two populations.

Conclusions: Our study replicated eight known CpG associations with T2D or BMI in an Arab population. Heterogeneity of the effects at all loci except CPT1A between the Qatari and UK studies suggests that the underlying mechanisms might depend on genetic background and environmental pressure. Our EWAS results provide a basis for comparison with other ethnicities.

Fig. 1

Boxplot of methylation beta values at cg19693031 (TXNIP) against diabetes state. The middle lines show the medians of the data while the boxes show the 25th to 75th percentiles (Q1 and Q3). The whiskers extend to include 99 % of the data above while circles represent outliers. Beta value distributions at this probe in the diabetics and healthy individuals showed a difference in the levels of background methylation (Wilcoxon test p value = 2.10 × 10⁻³)

Fig. 2

Scatter plots of BMI against methylation beta values. Red lines represent the slopes of the regression model. BMI values were corrected for age and sex