Epigenetic changes contributing to obesity and type 2 diabetes
Contact persons: Dr. Meriem Ouni, Leona Kovac, Markus Jähnert, Dr. Pascal Gottmann
Funding: German Center for Diabetes Research (DZD)
In addition to genetic causes, epigenetic differences such as changes in DNA methylation and the expression of non-coding RNA species (e.g. micro-RNAs) also contribute to obesity and type 2 diabetes. To identify differences in DNA methylation that are involved in increasing body weight and developing insulin resistance, we analyze the expression and DNA methylation profile of genetically identical mice (either B6 or NZO) that differ in their phenotype. Two important genes (DPP4, IGFBP2) were identified in B6 mice to be affected by the epigenome and to contribute to the development of fatty liver. In cooperation with colleagues from the DZD in Tübingen as well as in Lille and Malmö we observed similar changes in humans. DPP4 is increased in the liver of obese susceptible B6 mice due to a reduced DNA methylation. The phenotype of the liver-specific DPP4 transgenic mouse showed that higher DPP4 levels in the liver actually increase hepatic fat storage, trigger insulin resistance and increase body weight. DPP4 activity was also significantly increased in the plasma of NAFLD patients.
The situation is exactly the opposite for IGFBP2: Here the increased DNA methylation associates with a reduced expression of IGFBP2 in the liver of humans and mice, even before fat storage in the liver is elevated. In cooperation with the Department of Molecular Epidemiology, we tested if changed in plasma IGFBP2 levels and DNA methylation of IGFBP2 gene in blood cells differ between healthy people and those at risk of diabetes. A random cohort of 2,500 participants was generated from the EPIC Potsdam study, 820 of whom later developed type 2 diabetes. The IGFBP2 plasma levels of all persons were measured and the DNA methylation of the IGFBP2 gene was determined from 300 controls and 300 incident type 2 diabetics. We observed a clear inverse association between the circulating IGFBP2 concentrations and the type 2 diabetes risk. In addition, the methylation levels of the IGFBP2 gene for seven CpG positions were strongly correlated with the incidence of type 2 diabetes. It can therefore be assumed that the inhibition of the IGFBP2 gene promotes the development of type 2 diabetes.
In a current study, we are comparing the differences in DNA methylation and gene expression of islets of Langerhans from diabetes-susceptible and diabetes-resistant female NZO mice and compare the results with the corresponding patterns in islets of healthy people and type 2 diabetic patients. In addition, conserved methylation sites - as described above for the IGFBP2 gene – are evaluated in blood cells of EPIC participants in order to test if they improve the prediction of a diabetes risk.
The skeletal muscle plays a crucial role in the regulation of blood glucose and its insulin sensitivity which is severely impaired in obesity. In collaboration with DZD colleagues in Düsseldorf, we investigated which changes in gene expression and DNA methylation in skeletal muscle of severely obese people are triggered by bariatric (metabolic) surgery. Over a period of 2 weeks to 12 months, we observed the expected significant weight loss early on and an improved insulin sensitivity after about 3 months. Changes in gene expression that were associated with those of the epigenome were only detectable after 12 months. Genes involved in fat metabolism and mitochondrial function were particularly affected. We are currently investigating whether obese individuals with type 2 diabetes show similar changes in expression and DNA methylation patterns after bariatric surgery and to what they have reached the state of lean people with healthy metabolism It is also of great interest which other interventions have similar effects.
Micro-RNAs (miRNAs) are small non-coding RNAs with a length of 19-24 nucleotides. Usually, they inhibit the expression of several genes by complementarily binding to the 3 ’untranslated region (3’-UTR) of the mRNA of target genes and either inhibiting their translation or inducing the degradation of the mRNA. In a bioinformatics approach integrating QTL data, transcriptome data and target gene predictions, we screened for single nucleotide polymorphisms (SNPs) that theoretically disrupt the binding sites of miRNAs to target mRNAs and thus influence their expression. Of 566 genes that are located in obesity or diabetes QTL, have SNPs in miRNA-mRNA binding sites, 51 carry a conserved miRNA-mRNA binding site in the human genome. Results from corresponding knockout mice show that they can actually be linked to impaired metabolism. The deletion of 38 of the 51 genes in mice has already been documented and generated phenotypes that can be associated with metabolic disorders.
In islets of NZO and B6-ob/ob mice, we identified 94 differentially expressed miRNAs, of which 11 are located in diabetes QTL. Among these is miR-205-5p which exhibits the strongest expression difference. According to target prediction analyses, miR-205-5p affects genes involved in Wnt and calcium signalling as well as insulin secretion. In fact, overexpression of miR-205-5p in the insulinoma cell line INS-1 increased insulin expression, left-shifted the glucose-dependence of insulin secretion and suppressed the expression of the diabetes gene TCF7L2.