Chronological Age Interacts with the Circadian Melatonin Receptor 1B Gene Variation, Determining Fasting Glucose Concentrations in Mediterranean Populations. Additional Analyses on Type-2 Diabetes Risk

Abstract

Gene-age interactions have not been systematically investigated on metabolic phenotypes and this modulation will be key for a better understanding of the temporal regulation in nutrigenomics. Taking into account that aging is typically associated with both impairment of the circadian system and a decrease in melatonin secretion, we focused on the melatonin receptor 1B (MTNR1B)-rs10830963 C>G variant that has been associated with fasting glucose concentrations, gestational diabetes, and type-2 diabetes. Therefore, our main aim was to investigate whether the association between the MTNR1B-rs10830963 polymorphism and fasting glucose is age dependent. Our secondary aims were to analyze the polymorphism association with type-2 diabetes and explore the gene-pregnancies interactions on the later type-2 diabetes risk. Three Mediterranean cohorts (n = 2823) were analyzed. First, a cross-sectional study in the discovery cohort consisting of 1378 participants (aged 18 to 80 years; mean age 41 years) from the general population was carried out. To validate and extend the results, two replication cohorts consisting of elderly individuals were studied. In the discovery cohort, we observed a strong gene-age interaction (p = 0.001), determining fasting glucose in such a way that the increasing effect of the risk G-allele was much greater in young (p = 5.9 × 10^-10) than in elderly participants (p = 0.805). Consistently, the association of the MTNR1B-rs10830963 polymorphism with fasting glucose concentrations in the two replication cohorts (mean age over 65 years) did not reach statistical significance (p > 0.05 for both). However, in the elderly cohorts, significant associations between the polymorphism and type-2 diabetes at baseline were found. Moreover, in one of the cohorts, we obtained a statistically significant interaction between the MTNR1B polymorphism and the number of pregnancies, retrospectively assessed, on the type-2 diabetes risk. In conclusion, the association of the MTNR1B-rs10830963 polymorphism with fasting glucose is age-dependent, having a greater effect in younger people. However, in elderly subjects, associations of the polymorphism with type-2 diabetes were observed and our exploratory analysis suggested a modulatory effect of the number of past pregnancies on the future type-2 diabetes genetic risk.

Keywords: MTNR1B polymorphism; Mediterranean population; age-interaction; fasting glucose; gestational diabetes; heterogeneity; melatonin receptor; pregnancy; type-2 diabetes; women.

Figure 1

Associations between the MTNR1B-rs10830963 polymorphism and fasting glucose concentrations in: (A): whole population in the discovery cohort (OBENUTIC study) (n = 1378); (B): non-diabetic subjects (n = 1325) of the discovery cohort (OBENUTIC study). Means were adjusted for sex, age (as continuous), type-2 diabetes, and obesity. Variables are represented as means and SE. MTNR1B: melatonin receptor 1B gene; CC, CG and GG are the MTNR1B-rs10830963 genotypes. The p-values were obtained for the MTNR1B-rs10830963 polymorphism in the adjusted multivariable linear regression models, using two genetic models: p₁ as co-dominant model; p₂ as additive model.

Figure 2

Interaction between the MTNR1B-rs10830963 polymorphism and age (as a continuous variable) in determining fasting glucose concentrations in non-diabetic subjects from the discovery cohort (OBENUTIC study) (n = 1325). MTNR1B: melatonin receptor 1B gene; CC, CG and GG are the MTNR1B-rs10830963 genotypes. Results obtained in a multivariable linear regression model. Model and p-value for the interaction term were adjusted for sex, age (as continuous), and obesity. p-interaction = 0.009 in the hierarchical multivariate adjusted model. This continuous interaction was depicted by computing the predicted fasting glucose values (standardized) for each individual from the multivariate adjusted regression model and plotting them against chronological age (in years) by MTNR1B-rs10830963 genotypes.