Abstract

Objective. To investigate whether the effects of common genetic variants associated with fasting glucose in adults are detectable in healthy children.

Methods. Single nucleotide polymorphisms in MTNR1B (rs10830963), G6PC2 (rs560887) and GCK (rs4607517) were genotyped in 2,025 healthy European children aged 9-11 and 14-16 years. Associations with fasting glucose (FG), insulin, HOMA-IR and HOMA-B were investigated along with those observed for type 2 diabetes (T2D) variants available in this study (CDKN2A/B, IGF2BP2, CDKAL1, SLC30A8, HHEX-IDE, Chr 11p12).

Results Strongest associations were observed for G6PC2 and MTNR1B, with FG levels (95% CI) being 0.084 (0.06; 0.11) mmol/L, p=7.9×10⁻¹¹ and 0.069 (0.04; 0.09) mmol/L, p=1.9×10⁻⁷ higher per risk allele copy. A similar, but weaker trend was observed for GCK (0.028 (−0.006; 0.06) mmol/L, p=0.11). All three variants were associated with lower β-cell function (HOMA-B p=9.38x10⁻⁵, 0.004 and 0.04, respectively). SLC30A8 (rs13266634) was the only T2D variant associated with higher FG (0.033 mmol/L (0.01; 0.06); p=0.01). Calculating a genetic predisposition score adding the number of risk alleles of G6PC2, MTNR1B, GCK and SLC30A8 showed that glucose levels were successively higher in children carrying a greater number of risk alleles (p=7.1x10⁻¹⁷), with mean levels of 5.34 versus 4.91 mmol/L comparing children with 7 alleles (0.6% of all children) to those with none (0.5%). No associations were found for fasting insulin or HOMA-IR with any of the variants.

Conclusions. The effects of common polymorphisms influencing FG are apparent in healthy children, where the presence of multiple risk alleles amounts to a difference of more than a standard deviation of FG.