In This Issue of Diabetes

An epigenome-wide association study by Cardona et al. (p. 2315) has expanded the number of methylation variable positions (MVPs) in whole blood that appear to be associated with incident type 2 diabetes. They found that it was possible to confirm most of the associations in two independent studies involving different settings and ethnicities and that for at least one of the positions, modeling identified plausible evidence to implicate it in the development of type 2 diabetes. However, for the majority of the MVPs identified, adjustment for characteristics such as BMI or HbA_1c attenuated the associations. They say despite this, the associations are informative in terms of implicating genes and biological pathways that might be involved in the very early stages of type 2 diabetes development. Based on the European Prospective Investigation into Cancer and Nutrition (EPIC)-Norfolk cohort, the authors identified an incident type 2 diabetes case-cohort and then used whole blood samples collected at baseline and up to 11 years prior to type 2 diabetes onset for the investigation. With methylation array profiling, they managed to identify 15 novel MVPs that were associated with incident type 2 diabetes and also confirm 3 others previously identified in other studies. All of the MVPs identified in the EPIC-Norfolk cohort were directionally consistent in the two further independent cohorts, effectively confirming their association with type 2 diabetes. They go on to systematically investigate the MVPs for direct causal effects in type 2 diabetes but eventually rule out most via Mendelian randomization analyses. They do however identify one particular candidate, cg00574958 at CPT1A, as having a possible direct causal role in type 2 diabetes. Commenting further, author Ken K. Ong told us: “This work shows that there are many robust and replicable epigenetic changes in blood cells that precede the onset of type 2 diabetes. While a minority of these changes might indicate pathways leading to diabetes, the findings also help us understand how the body responds to high glucose levels.”

Cardona et al. Epigenome-wide association study of incident type 2 diabetes in a British population: EPIC-Norfolk study. Diabetes 2019;68:2315–2326

A set of potentially renoprotective genes may be upregulated at early stages of diabetic nephropathy but downregulated at more advanced stages of the disease, according to Fan et al. (p. 2301). Specifically, they point toward the genes in the retinoic acid pathway and also the glucagon-like peptide 1 receptor (GLP1R), which are known to have renal protective effects. The findings come from RNA sequencing of whole-kidney biopsy samples of patients with either early (n = 6) or advanced (n = 22) diabetic nephropathy. For controls (n = 9), the authors also included patient samples from the unaffected portion of tumor nephrectomies. They found that overall gene expression profiles separated into three distinct clusters according to control or early/advanced diabetic nephropathy status (see principle component analysis output on right). When they compared early diabetic nephropathy versus control participants, a fairly consistent pattern of differential gene expression relating to immune response and inflammation emerged. However, when they compared early versus advanced diabetic nephropathy, they found differences in gene expression that indicated inflammation as a key process for disease progression. Looking at correlations between kidney function (estimated glomerular filtration rate), histological scores, and gene expression, unique pathways appear associated with the decline of functions and histological features. Using a deconvolution analysis, the authors were able to estimate the relative fractions of different cell types present in the samples according to disease status. Finally, by using immunostaining, they confirmed their findings related to retinoic acid and GLP1R as well as the infiltration of immune cells. Discussing the findings, they defend their use of whole-kidney samples to minimize the artificial effects of tissue manipulation on gene expression. They go on to suggest the data will help the wider community further understand the molecular mechanisms and pathogenesis of diabetic nephropathy. Commenting further, author John Cijiang He told us: “The major impact of this study is the discovery of early protective mechanisms against progression of diabetic nephropathy, and therefore we could develop early preventive measures for these patients.”

Fan et al. Comparison of kidney transcriptomic profiles of early and advanced diabetic nephropathy reveals potential new mechanisms for disease progression. Diabetes 2019;68:2301–2315

Using a genome-wide association study approach, Li-Gao et al. (p. 2327) describe how a genetic variant termed rs505922:C correlates with early-phase insulin response. The finding potentially explains why carriers of the allele have increased type 2 diabetes risk that might be explained by decreased early-phase insulin secretion. The conclusions come from an analysis of the Netherlands Epidemiology of Obesity (NEO) study that involved thousands of individuals undergoing assessments for early-phase insulin response to a liquid mixed meal. Blood samples were then assessed in a discovery and replication genome-wide analysis in an attempt to identify genetic signals that might correlate with insulin response. After applying strict conditions to account for false discoveries, they identified a series of hits located in the ABO gene. Notably, the rs505922:C allele was the only one that could be replicated in a second cohort. They point out that the ABO gene is linked to blood type and that individuals with the O-type had a lower fasting level of insulin but larger increases following the mixed meal compared to non-O types. Additional in vitro experiments with a murine pancreatic β-cell line showed that knockout of the ABO gene resulted in a 60% reduction its expression. Also, after glucose stimulation, the increase in insulin secretion was lower in the cells with the knockout compared to cells without it. Commenting further, author Ruifang Li-Gao told us: “ABO gene is ubiquitously expressed and encodes a glycosyltransferase, an enzyme that adds N-acetylgalactosamine to many proteins. This explains the wide variety of traits that have also been associated with ABO blood type, such as venous thromboembolic disease and hematocrit percentage. Our study showed that a lifelong difference in response to meals may result in a true discrepancy with regard to glucose homeostasis and subsequent increased risk of type 2 diabetes. ABO gene also plays a role in explaining the difference of meal response among individuals.”

Li-Gao et al. Genome-wide association study on the early phase insulin response to a liquid mixed meal: results from the NEO study. Diabetes 2019;68:2327–2336

The roles of the transcription factor Neurod1 in mice and NEUROD1 in human cells in the development of β-cells are explored by Romer et al. (p. 2259). They clarify that Neurod1 appears to be important for proliferation of β-cells in mice, but, in contrast, NEUROD1 might play an essential role in the differentiation and maturation of fully functioning insulin-expressing β-cells in humans. While the findings provide insight into mechanisms involved in the development and function of β-cells, the authors suggest that they also add to evidence that mutations that inactivate NEUROD1 might be the root cause of permanent neonatal diabetes. Based on experiments using mouse models and also human embryonic stem cells (HESCs), the authors look at the effects of a knockout of either the mouse gene Neurod1 or disruption of NEUROD1 in HESCs. They found that perinatal reduction of α- and β-cells in the pancreas of Neurod1 knockout mice is primarily due to defects in proliferative expansion, with only minor contribution from apoptosis. They also found that the Neurod1 knockout changed the regulation of a series of genes involved in β-cell maturation, cell cycle progression, and insulin secretion as well as the control of hormone secretion and levels. Moving to HESCs, they use a protocol to convert the cells to insulin-producing β-cells and show that deletion of NEUROD1 severely impaired their differentiation to β-cells but that survival and proliferation was not affected. They also show that NEUROD1 is required to activate a transcription network for the full development of β-cells that produce insulin. Commenting further, author Lori Sussel told us: “Although it appears there are differences in NEUROD1 function between mice and humans, we are gratified that many of the gene networks regulated by NEUROD1 are conserved in both species. We hope that these studies will begin to shed light on the pancreatic defects associated in NEUROD1 mutations in humans.”

Romer et al. Murine perinatal β-cell proliferation and the differentiation of human stem cell–derived insulin-expressing cells require NEUROD1. Diabetes 2019;68:2259–2271