Melatonin receptor 1 B polymorphisms associated with the risk of gestational diabetes mellitus

The prevalence of type 2 diabetes mellitus (T2DM) in Korean population has dramatically increased over last decades. Although Asian populations traditionally had a low percentage of T2DM patients, it has increased drastically in recent decades. This is largely due to the fact that Asian countries are adopting western lifestyle and diets. However, a recent discovery of diabetes-susceptible loci on human chromosomes suggest that genetic factors may also play a role in the disease development [1].

Gestational diabetes mellitus (GDM) is a condition in which pregnant women exhibit glucose intolerance in various degrees [2], affecting approximately 2-14% of pregnancies [1, 3, 4]. Women with GDM show similar physiological and genetic characteristics found in diabetes outside of pregnancy, and not surprisingly, women with GDM possess higher risk for developing T2DM when they are not pregnant. Therefore, studying GDM is a good way to study early pathogenesis of diabetes and possibly develop treatment or remedy for the disease [5]. However, while genetic studies on T2DM are very robust [6, 7], there are relatively fewer genetic studies for GDM.

Advance of technology in the genetics field has provided us with a number of useful tools to study human genome. Among them, genome wide association studies (GWAS) are a powerful and useful way to detect genes associated with various diseases, including diabetes. Recently, a couple of studies have revealed that the genetic variants in melatonin receptor 1 B (MTNR1B) gene are associated with T2DM and fasting glucose levels [8, 9]. MTNR1B gene encodes MT2 protein which, along with MT1 protein encoded by MTNR1A, is one of two high-affinity forms of melatonin receptor. This gene product is also an integral membrane protein forming a G-protein coupled 7-transmembrane receptor. Melatonin, also known chemically as N-acetyl-5-methoxytryptamine, is a primary neurohormone secreted by pineal gland. It is mostly found in retina and brain, and its main function is thought to be the regulation of circadian rhythm by translating photoperiodic information from the eyes to the brain. There have been some studies suggesting that insulin level is regulated by circadian clock. Furthermore, T2DM patients have exhibited impaired melatonin secretion and circadian rhythm [10].

To date, several studies have shown the association between MTNR1B and T2DM [6, 11], but there has yet to be a study which looked into the association between MTNR1B and GDM. Since GDM shares many clinical features with T2DM, there is a high possibility that MTNR1B is associated with predisposition of GDM. Therefore, we conducted an association study between two polymorphisms of MTNR1B, which were previously associated with T2DM, using 928 GDM patients and 990 controls. In addition, we also examined the possible association between the two SNPs of MTNR1B and clinical phenotypes related to GDM, such as insulin sensitivity and beta-cell function.

Nine hundred and twenty eight GDM patients were recruited for the present study, and we also recruited 990 pregnant women with normal glucose tolerance as controls. The clinical profiles of the subjects are summarized in Table 1, with characteristics related with T2DM such as the area under glucose curve (AUC-G), fasting plasma insulin (FPI), and fasting plasma glucose (FPG). We also obtained homeostatic model assessment data for both groups in beta-cell function and insulin resistance (HOMA-B and HOMA-IR, respectively). Most of the phenotypes investigated for the subjects showed significant difference between the GDM patients group and the control group, (Table 1, P value < 0.0001 for all phenotypes except mean gestational week), which was to be expected because the phenotypes that showed the significant differences were associated with the diabetic condition. GDM patients were older, possessed higher BMI than the NGT women and clearly exhibited the clinical characteristics of T2DM, as shown in Table 1; in comparison to NGT women, GDM patients exhibited higher blood glucose levels and lower beta-cell function when insulin resistance was increased.

In previous studies on the rs1387153 and rs10830963, researchers found strong associations between the two SNPs of MTNR1B and T2DM, and also with FPG levels, which is an important phenotype for diabetes. A study in European population, which included French and Danish among other nations, showed that rs1387153 was significantly associated with FPG level (P = 1.3 × 10^-7, adjusted genome-wide P = 0.04) and T2DM (OR (95% CI) = 1.15(1.08-1.22), P = 6.3 × 10^-5). Another study in European population found significant associations of rs10830963 with FPG (P = 3.2 × 10^-50) and T2DM (OR (95% CI) = 1.09(1.05-1.12), P = 3.3 × 10^-7). The Two SNPs were closely related with each other, as evidenced by the linkage disequilibrium test (|D'| = 0.89) in our study. Their haplotype analyses results showed that ht1 (C/C) and ht2 (T/G) were mostly tagged by rs10830963 and rs1387153, respectively (>92%). Therefore, ht1/ht2 showed similar associations with each SNP, respectively.

Here, we performed the association studies in Korean pregnant women and we found significant associations between the SNPs and GDM, with enough samples for high statistical power. It is well known that T2DM and GDM are closely related diseases, since they exhibit similar characteristics such as glucose intolerance. However, there had yet to be a study that looked into the association between the polymorphisms of MTNR1B and GDM, and our study confirmed the relations between the two. Also, we carried out regression analyses between the polymorphisms of MTNR1B and various phenotypes including FPG. Although both SNPs showed associations with GDM and FPG, our results suggested that the two genetic variants of MTNR1B were stronger risk factors for GDM in Korean population compared to the previous results for T2DM in European population (OR (95% CI) = 1.44 (1.17-1.78) for rs1387153 in dominant inheritance model and 1.46 (1.18 - 1.81) for rs10830963 in the present study and OR = 1.15 (1.08-1.22) for rs1387153 and OR = 1.09 (1.05-1.12) for rs10830963 in the two previous studies). We suspect that the genetic differences between GDM and T2DM and the population difference between Europeans and Asians could have contributed to this result. Previously, there have been a few cases where a gene associated with T2DM was not associated with GDM at all [20], or showed different effect sizes [21, 22], and our results suggest that MTNR1B affects T2DM and GDM in varying degrees as well.

Moreover, our results suggest that the two polymorphisms investigated are associated with beta-cell function (Table 5). Association between beta-cell function and MTNR1B was previously reported [18, 23], which shows that we were able to replicate the result in Korean population, strengthening the notion that MTNR1B polymorphisms are related with impaired beta-cell function. Recently, several groups of scientists have studied the association between the gene variants of MTNR1B and glucose tolerances. rs10830963 was found to be associated with FPG and decreased beta-cell function in a group of obese children, which is consistent with our finding [24]. Three independent studies of the MTNR1B genetic variants in Han Chinese subjects also found significant associations for increased FPG, impaired beta-cell function, glycated hemoglobin, and T2DM [18, 25, 26]. Furthermore, a study with European populations also found significant associations between variations of MTNR1B with BMI and FPG, but not with maturity-onset diabetes of the young (MODY) or T2DM [27]. Although some of their results do not agree with each other in the association with T2DM, these findings firmly back up the association between the SNPs in our study and FPG or impaired beta-cell function. On the other hand, we could not find any significant associations between insulin resistance (HOMA-IR) and the two genetic variants. Since our results are backed with high statistical power, this leads us to conclude that two genetic variants of MTNR1B may be associated with the disease by affecting glucose metabolism through impaired insulin secretion, as previously suggested by other studies with KCNQ1 genetic variants [8, 18, 28].

In addition, further conditional analyses results suggested that rs10830963 might be more likely functional in case/control analysis, although not clear in GDM-related phenotype analyses.

Although our results showed the significant associations with GDM and several diabetic characteristics, there are a couple of limitations. First, our study only concentrated on pregnant women among Korean population, so we cannot conclude that the MTNR1B gene variants are associated with FPG or impaired beta-cell function in all Korean population. Also, even though our study strongly suggests that the SNPs may also be associated with T2DM in Korean population, this is not confirmed yet. Any further studies on these two genetic variants in Korean population should concentrate on these parts.

The present study showed that two MTNR1B polymorphisms were associated with increased risk for GDM in Korean female population. Two polymorphisms rs1387153 and rs10830963 also showed significant associations with FPG and beta-cell function, but not with insulin resistance. Further conditional analyses results suggested that rs10830963 might be more likely functional in case/control analysis, although not clear in GDM-related phenotype analyses. The effective sizes found between the two polymorphisms and FPG was stronger compared to previous studies, which is possibly due to the genetic difference between European and Korean populations, or the difference between GDM and T2DM. Based on the current results, we suspect that these two polymorphisms will have significant associations with increased risk of GDM in other populations as well. Also, our discovery would be helpful for understanding of genetic etiology of GDM as well.