GAD autoantibodies (GADAs) are sensitive markers of islet autoimmunity and type 1 diabetes. They form the basis of robust prediction models and are widely used for the recruitment of subjects at high risk of type 1 diabetes to prevention trials. However, GADAs are also found in many individuals at low risk of diabetes progression. To identify the sources of diabetes-irrelevant GADA reactivity, we analyzed data from the 2009 and 2010 Diabetes Autoantibody Standardization Program GADA workshop and found that binding of healthy control sera varied according to assay type. The characterization of control sera found positive by radiobinding assay (RBA), but negative by ELISA, showed that many of these sera reacted to epitopes in the N-terminal region of the molecule. This finding prompted development of an N-terminally truncated GAD65 radiolabel, (35)S-GAD65(96-585), which improved the performance of most GADA RBAs participating in an Islet Autoantibody Standardization Program GADA substudy. These detailed workshop comparisons have identified a source of disease-irrelevant signals in GADA RBAs and suggest that N-terminally truncated GAD labels will enable more specific measurement of GADAs in type 1 diabetes.

Observationally, low levels of HDL cholesterol are consistently associated with increased risk of type 2 diabetes. Therefore, plasma HDL cholesterol increasing has been suggested as a novel therapeutic option to reduce the risk of type 2 diabetes. Whether levels of HDL cholesterol are causally associated with type 2 diabetes is unknown. In a prospective study of the general population (n = 47,627), we tested whether HDL cholesterol-related genetic variants were associated with low HDL cholesterol levels and, in turn, with an increased risk of type 2 diabetes. HDL cholesterol-decreasing gene scores and allele numbers associated with up to -13 and -20% reductions in HDL cholesterol levels. The corresponding theoretically predicted hazard ratios for type 2 diabetes were 1.44 (95% CI 1.38-1.52) and 1.77 (1.61-1.95), whereas the genetic estimates were nonsignificant. Genetic risk ratios for type 2 diabetes for a 0.2 mmol/L reduction in HDL cholesterol were 0.91 (0.75-1.09) and 0.93 (0.78-1.11) for HDL cholesterol-decreasing gene scores and allele numbers, respectively, compared with the corresponding observational hazard ratio of 1.37 (1.32-1.42). In conclusion, genetically reduced HDL cholesterol does not associate with increased risk of type 2 diabetes, suggesting that the corresponding observational association is due to confounding and/or reverse causation.

Successful weight loss is variable for reasons not fully elucidated. Whether effective weight loss results from smaller reductions in energy expenditure during caloric restriction is not known. We analyzed whether obese individuals with a "thrifty" phenotype, that is, greater reductions in 24-h energy expenditure during fasting and smaller increases with overfeeding, lose less weight during caloric restriction than those with a "spendthrift" phenotype. During a weight-maintaining period, 24-h energy expenditure responses to fasting and 200% overfeeding were measured in a whole-room indirect calorimeter. Volunteers then underwent 6 weeks of 50% caloric restriction. We calculated the daily energy deficit (kilocalories per day) during caloric restriction, incorporating energy intake and waste, energy expenditure, and daily activity. We found that a smaller reduction in 24-h energy expenditure during fasting and a larger response to overfeeding predicted more weight loss over 6 weeks, even after accounting for age, sex, race, and baseline weight, as well as a greater rate of energy deficit accumulation. The success of dietary weight loss efforts is influenced by the energy expenditure response to caloric restriction. Greater decreases in energy expenditure during caloric restriction predict less weight loss, indicating the presence of thrifty and spendthrift phenotypes in obese humans.

Because regulatory T-cell (Treg) development can be induced by the same agonist self-antigens that induce negative selection, perturbation of apoptosis will affect both negative selection and Treg development. But how the processes of thymocyte deletion versus Treg differentiation bifurcate and their relative importance for tolerance have not been studied in spontaneous organ-specific autoimmune disease. We addressed these questions by removing a critical mediator of thymocyte deletion, BIM, in the NOD mouse model of autoimmune diabetes. Despite substantial defects in the deletion of autoreactive thymocytes, BIM-deficient NOD (NODBim(-/-)) mice developed less insulitis and were protected from diabetes. BIM deficiency did not impair effector T-cell function; however, NODBim(-/-) mice had increased numbers of Tregs, including those specific for proinsulin, in the thymus and peripheral lymphoid tissues. Increased levels of Nur77, CD5, GITR, and phosphorylated IκB-α in thymocytes from NODBim(-/-) mice suggest that autoreactive cells receiving strong T-cell receptor signals that would normally delete them escape apoptosis and are diverted into the Treg pathway. Paradoxically, in the NOD model, reduced thymic deletion ameliorates autoimmune diabetes by increasing Tregs. Thus, modulating apoptosis may be one of the ways to increase antigen-specific Tregs and prevent autoimmune disease.

Protein hyperacetylation is associated with glucose intolerance and insulin resistance, suggesting that the enzymes regulating the acetylome play a role in this pathological process. Sirtuin 3 (SIRT3), the primary mitochondrial deacetylase, has been linked to energy homeostasis. Thus, it is hypothesized that the dysregulation of the mitochondrial acetylation state, via genetic deletion of SIRT3, will amplify the deleterious effects of a high-fat diet (HFD). Hyperinsulinemic-euglycemic clamp experiments show, for the first time, that mice lacking SIRT3 exhibit increased insulin resistance due to defects in skeletal muscle glucose uptake. Permeabilized muscle fibers from HFD-fed SIRT3 knockout (KO) mice showed that tricarboxylic acid cycle substrate-based respiration is decreased while fatty acid-based respiration is increased, reflecting a fuel switch from glucose to fatty acids. Consistent with reduced muscle glucose uptake, hexokinase II (HKII) binding to the mitochondria is decreased in muscle from HFD-fed SIRT3 KO mice, suggesting decreased HKII activity. These results show that the absence of SIRT3 in HFD-fed mice causes profound impairments in insulin-stimulated muscle glucose uptake, creating an increased reliance on fatty acids. Insulin action was not impaired in the lean SIRT3 KO mice. This suggests that SIRT3 protects against dietary insulin resistance by facilitating glucose disposal and mitochondrial function.

The effects of dyslipidemia on the risk of type 2 diabetes (T2D) and related traits are not clear. We used regression models and 140 lipid-associated genetic variants to estimate associations between circulating HDL cholesterol (HDL-C), LDL cholesterol (LDL-C), and triglycerides and T2D and related traits. Each genetic test was corrected for effects of variants on the other two lipid types and surrogates of adiposity. We used the largest data sets available: 34,840 T2D case and 114,981 control subjects from the DIAGRAM (DIAbetes Genetics Replication And Meta-analysis) consortium and up to 133,010 individuals without diabetes for insulin secretion and sensitivity from the MAGIC (Meta-Analyses of Glucose and Insulin-related traits Consortium) and GENESIS (GENEticS of Insulin Sensitivity) studies. Eight of 21 associations between groups of variants and diabetes traits were significant at the nominal level, including those between genetically determined lower HDL-C (β = -0.12, P = 0.03) and T2D and genetically determined lower LDL-C (β = -0.21, P = 5 × 10(-6)) and T2D. Although some of these may represent causal associations, we discuss why caution must be used when using Mendelian randomization in the context of circulating lipid levels and diabetes traits. In conclusion, we found evidence of links between genetic variants associated with lipids and T2D, but deeper knowledge of the underlying genetic mechanisms of specific lipid variants is needed before drawing definite conclusions about causality based on Mendelian randomization methodology.

This study investigated how maternal overnutrition and obesity regulate expression and activation of proteins that facilitate lipid transport in the placenta. To create a maternal overnutrition and obesity model, primiparous C57BL/6 mice were fed a high-fat (HF) diet throughout gestation. Fetuses from HF-fed dams had significantly increased serum levels of free fatty acid and body fat. Despite no significant difference in placental weight, lipoprotein lipase (LPL) protein levels and activity were remarkably elevated in placentas from HF-fed dams. Increased triglyceride content and mRNA levels of CD36, VLDLr, FABP3, FABPpm, and GPAT2 and -3 were also found in placentas from HF-fed dams. Although both peroxisome proliferator-activated receptor-γ (PPARγ) and CCAAT/enhancer binding protein-α protein levels were significantly increased in placentas of the HF group, only PPARγ exhibited a stimulative effect on LPL expression in cultured JEG-3 human trophoblasts. Maternal HF feeding remarkably decreased SIRT1 expression in placentas. Through use of an SIRT1 activator and inhibitor and cultured trophoblasts, an inhibitory effect of SIRT1 on LPL expression was demonstrated. We also found that SIRT1 suppresses PPARγ expression in trophoblasts. Most importantly, inhibition of PPARγ abolished the SIRT1-mediated regulatory effect on LPL expression. Together, these results indicate that maternal overnutrition induces LPL expression in trophoblasts by reducing the inhibitory effect of SIRT1 on PPARγ.

Production of adrenomedullin (ADM), a vasodilator peptide, increases in response to ischemia and hypoxia in the vascular wall and the kidney. This may be an adaptive response providing protection against organ damage. We investigated the hypothesis that ADM has a nephroprotective effect in two prospective cohorts of patients with type 2 diabetes recruited in France. The highest tertile of plasma MR-proADM (a surrogate for ADM) concentration at baseline was associated with the risk of renal outcomes (doubling of plasma creatinine concentration and/or progression to end-stage renal disease) during follow-up in both cohorts. Four SNPs in the ADM gene region were associated with plasma MR-proADM concentration at baseline and with eGFR during follow-up in both cohorts. The alleles associated with lower eGFR were also associated with lower plasma MR-proADM level. In conclusion, plasma MR-proADM concentration was associated with renal outcome in patients with type 2 diabetes. Our data suggest that the ADM gene modulates the genetic susceptibility to nephropathy progression. Results are consistent with the hypothesis of a reactive rise of ADM in diabetic nephropathy, blunted in risk alleles carriers, and with a nephroprotective effect of ADM. A possible therapeutic effect of ADM receptor agonists in diabetic renal disease would be worth investigating.

We investigated whether circulating TGF-β1-regulated miRNAs detectable in plasma are associated with the risk of rapid progression to end-stage renal disease (ESRD) in a cohort of proteinuric patients with type 1 diabetes (T1D) and normal eGFR. Plasma specimens obtained at entry to the study were examined in two prospective subgroups that were followed for 7-20 years (rapid progressors and nonprogressors), as well as a reference panel of normoalbuminuric T1D patients. Of the five miRNAs examined in this study, let-7c-5p and miR-29a-3p were significantly associated with protection against rapid progression and let-7b-5p and miR-21-5p were significantly associated with the increased risk of ESRD. In logistic analysis, controlling for HbA1c and other covariates, let-7c-5p and miR-29a-3p were associated with more than a 50% reduction in the risk of rapid progression (P ≤ 0.001), while let-7b-5p and miR-21-5p were associated with a >2.5-fold increase in the risk of ESRD (P ≤ 0.005). This study is the first prospective study to demonstrate that circulating TGF-β1-regulated miRNAs are deregulated early in T1D patients who are at risk for rapid progression to ESRD.

Diffuse congenital hyperinsulinism in infancy (CHI-D) arises from mutations inactivating the KATP channel; however, the phenotype is difficult to explain from electrophysiology alone. Here we studied wider abnormalities in the β-cell and other pancreatic lineages. Islets were disorganized in CHI-D compared with controls. PAX4 and ARX expression was decreased. A tendency toward increased NKX2.2 expression was consistent with its detection in two-thirds of CHI-D δ-cell nuclei, similar to the fetal pancreas, and implied immature δ-cell function. CHI-D δ-cells also comprised 10% of cells displaying nucleomegaly. In CHI-D, increased proliferation was most elevated in duct (5- to 11-fold) and acinar (7- to 47-fold) lineages. Increased β-cell proliferation observed in some cases was offset by an increase in apoptosis; this is in keeping with no difference in INSULIN expression or surface area stained for insulin between CHI-D and control pancreas. However, nuclear localization of CDK6 and P27 was markedly enhanced in CHI-D β-cells compared with cytoplasmic localization in control cells. These combined data support normal β-cell mass in CHI-D, but with G1/S molecules positioned in favor of cell cycle progression. New molecular abnormalities in δ-cells and marked proliferative increases in other pancreatic lineages indicate CHI-D is not solely a β-cell disorder.

Understanding distinct gene expression patterns of normal adult and developing fetal human pancreatic α- and β-cells is crucial for developing stem cell therapies, islet regeneration strategies, and therapies designed to increase β-cell function in patients with diabetes (type 1 or 2). Toward that end, we have developed methods to highly purify α-, β-, and δ-cells from human fetal and adult pancreata by intracellular staining for the cell-specific hormone content, sorting the subpopulations by flow cytometry, and, using next-generation RNA sequencing, we report the detailed transcriptomes of fetal and adult α- and β-cells. We observed that human islet composition was not influenced by age, sex, or BMI, and transcripts for inflammatory gene products were noted in fetal β-cells. In addition, within highly purified adult glucagon-expressing α-cells, we observed surprisingly high insulin mRNA expression, but not insulin protein expression. This transcriptome analysis from highly purified islet α- and β-cell subsets from fetal and adult pancreata offers clear implications for strategies that seek to increase insulin expression in type 1 and type 2 diabetes.

The endocrine role of the skeleton in regulating energy metabolism is supported by a feed-forward loop between circulating osteoblast (OB)-derived undercarboxylated osteocalcin (Glu-OCN) and pancreatic β-cell insulin; in turn, insulin favors osteocalcin (OCN) bioactivity. These data suggest the existence of a negative regulation of this cross talk between OCN and insulin. Recently, we identified delta like-1 (DLK1) as an endocrine regulator of bone turnover. Because DLK1 is colocalized with insulin in pancreatic β-cells, we examined the role of DLK1 in insulin signaling in OBs and energy metabolism. We show that Glu-OCN specifically stimulates Dlk1 expression by the pancreas. Conversely, Dlk1-deficient (Dlk1(-/-) ) mice exhibited increased circulating Glu-OCN levels and increased insulin sensitivity, whereas mice overexpressing Dlk1 in OB displayed reduced insulin secretion and sensitivity due to impaired insulin signaling in OB and lowered Glu-OCN serum levels. Furthermore, Dlk1(-/-) mice treated with Glu-OC experienced significantly lower blood glucose levels than Glu-OCN-treated wild-type mice. The data suggest that Glu-OCN-controlled production of DLK1 by pancreatic β-cells acts as a negative feedback mechanism to counteract the stimulatory effects of insulin on OB production of Glu-OCN, a potential mechanism preventing OCN-induced hypoglycemia.

Loss of pancreatic islet β-cell mass and β-cell dysfunction are central in the development of type 2 diabetes (T2DM). We recently showed that mature human insulin-containing β-cells can convert into glucagon-containing α-cells ex vivo. This loss of β-cell identity was characterized by the presence of β-cell transcription factors (Nkx6.1, Pdx1) in glucagon(+) cells. Here, we investigated whether the loss of β-cell identity also occurs in vivo, and whether it is related to the presence of (pre)diabetes in humans and nonhuman primates. We observed an eight times increased frequency of insulin(+) cells coexpressing glucagon in donors with diabetes. Up to 5% of the cells that were Nkx6.1(+) but insulin(-) coexpressed glucagon, which represents a five times increased frequency compared with the control group. This increase in bihormonal and Nkx6.1(+)glucagon(+)insulin(-) cells was also found in islets of diabetic macaques. The higher proportion of bihormonal cells and Nkx6.1(+)glucagon(+)insulin(-) cells in macaques and humans with diabetes was correlated with the presence and extent of islet amyloidosis. These data indicate that the loss of β-cell identity occurs in T2DM and could contribute to the decrease of functional β-cell mass. Maintenance of β-cell identity is a potential novel strategy to preserve β-cell function in diabetes.

Pancreatic β-cells are critical in the regulation of glucose homeostasis by controlled secretion of insulin in mammals. Activation of protein kinase A by cAMP is shown to be responsible for enhancing this pathway, which is countered by phosphodiesterase (PDE) that converts cAMP to AMP and turns off the signal. Salt-inducible kinases (SIKs) were also known to inhibit cAMP signaling, mostly by promoting inhibitory phosphorylation on CREB-regulated transcription coactivators. Here, we showed that SIK1 regulates insulin secretion in β-cells by modulating PDE4D and cAMP concentrations. Haploinsufficiency of SIK1 led to the improved glucose tolerance due to the increased glucose-stimulated insulin secretion. Depletion of SIK1 promoted higher cAMP concentration and increased insulin secretion from primary islets, suggesting that SIK1 controls insulin secretion through the regulation of cAMP signaling. By using a consensus phosphorylation site of SIK1, we identified PDE4D as a new substrate for this kinase family. In vitro kinase assay as well as mass spectrometry analysis revealed that the predicted Ser(136) and the adjacent Ser(141) of PDE4D are critical in SIK1-mediated phosphorylation. We found that overexpression of either SIK1 or PDE4D in β-cells reduced insulin secretion, while inhibition of PDE4 activity by rolipram or knockdown of PDE4D restored it, showing indeed that SIK1-dependent phosphorylation of PDE4D is critical in reducing cAMP concentration and insulin secretion from β-cells. Taken together, we propose that SIK1 serves as a part of a self-regulatory circuit to modulate insulin secretion from pancreatic β-cells by controlling cAMP concentration through modulation of PDE4D activity.

The first signs of autoimmune activation leading to β-cell destruction in type 1 diabetes (T1D) appear during the first months of life. Thus, the perinatal period offers a suitable time window for disease prevention. Moreover, thymic selection of autoreactive T cells is most active during this period, providing a therapeutic opportunity not exploited to date. We therefore devised a strategy by which the T1D-triggering antigen preproinsulin fused with the immunoglobulin (Ig)G Fc fragment (PPI-Fc) is delivered to fetuses through the neonatal Fc receptor (FcRn) pathway, which physiologically transfers maternal IgGs through the placenta. PPI-Fc administered to pregnant PPIB15-23 T-cell receptor-transgenic mice efficiently accumulated in fetuses through the placental FcRn and protected them from subsequent diabetes development. Protection relied on ferrying of PPI-Fc to the thymus by migratory dendritic cells and resulted in a rise in thymic-derived CD4(+) regulatory T cells expressing transforming growth factor-β and in increased effector CD8(+) T cells displaying impaired cytotoxicity. Moreover, polyclonal splenocytes from nonobese diabetic (NOD) mice transplacentally treated with PPI-Fc were less diabetogenic upon transfer into NOD.scid recipients. Transplacental antigen vaccination provides a novel strategy for early T1D prevention and, further, is applicable to other immune-mediated conditions.

The homeodomain transcription factor Pdx1 controls pancreas organogenesis, specification of endocrine pancreas progenitors, and the postnatal growth and function of pancreatic β-cells. Pdx1 expression in human-derived stem cells is used as a marker for induced pancreatic precursor cells. Unfortunately, the differentiation efficiency of human pancreatic progenitors into functional β-cells is poor. In order to gain insight into the genes that Pdx1 regulates during differentiation, we performed Pdx1 chromatin immunoprecipitation followed by high-throughput sequencing of embryonic day (e) 13.5 and 15.5 mouse pancreata. From this, we identified the transcription factor Teashirt zinc finger 1 (Tshz1) as a direct Pdx1 target. Tshz1 is expressed in developing and adult insulin- and glucagon-positive cells. Endocrine cells are properly specified in Tshz1-null embryos, but critical regulators of β-cell (Pdx1 and Nkx6.1) and α-cell (MafB and Arx) formation and function are downregulated. Adult Tshz1(+/-) mice display glucose intolerance due to defects in glucose-stimulated insulin secretion associated with reduced Pdx1 and Clec16a expression in Tshz1(+/-) islets. Lastly, we demonstrate that TSHZ1 levels are reduced in human islets of donors with type 2 diabetes. Thus, we position Tshz1 in the transcriptional network of maturing β-cells and suggest that its dysregulation could contribute to the islet phenotype of human type 2 diabetes.

Toll-like receptors (TLRs) have been implicated in the pathogenesis of type 2 diabetes. We examined the function of TLR3 in glucose metabolism and type 2 diabetes-related phenotypes in animals and humans. TLR3 is highly expressed in the pancreas, suggesting that it can influence metabolism. Using a diet-induced obesity model, we show that TLR3-deficient mice had enhanced glycemic control, facilitated by elevated insulin secretion. Despite having high insulin levels, Tlr3(-/-) mice did not experience disturbances in whole-body insulin sensitivity, suggesting that they have a robust metabolic system that manages increased insulin secretion. Increase in insulin secretion was associated with upregulation of islet glucose phosphorylation as well as exocytotic protein VAMP-2 in Tlr3(-/-) islets. TLR3 deficiency also modified the plasma lipid profile, decreasing VLDL levels due to decreased triglyceride biosynthesis. Moreover, a meta-analysis of two healthy human populations showed that a missense single nucleotide polymorphism in TLR3 (encoding L412F) was linked to elevated insulin levels, consistent with our experimental findings. In conclusion, our results increase the understanding of the function of innate receptors in metabolic disorders and implicate TLR3 as a key control system in metabolic regulation.

We aimed to investigate the causal effect of circulating uric acid concentrations on type 2 diabetes risk. A Mendelian randomization study was performed using a genetic score with 24 uric acid-associated loci. We used data of the European Prospective Investigation into Cancer and Nutrition (EPIC)-InterAct case-cohort study, comprising 24,265 individuals of European ancestry from eight European countries. During a mean (SD) follow-up of 10 (4) years, 10,576 verified incident case subjects with type 2 diabetes were ascertained. Higher uric acid was associated with a higher diabetes risk after adjustment for confounders, with a hazard ratio (HR) of 1.20 (95% CI 1.11, 1.30) per 59.48 µmol/L (1 mg/dL) uric acid. The genetic score raised uric acid by 17 µmol/L (95% CI 15, 18) per SD increase and explained 4% of uric acid variation. By using the genetic score to estimate the unconfounded effect, we found that a 59.48 µmol/L higher uric acid concentration did not have a causal effect on diabetes (HR 1.01 [95% CI 0.87, 1.16]). Including data from the Diabetes Genetics Replication And Meta-analysis (DIAGRAM) consortium, increasing our dataset to 41,508 case subjects with diabetes, the summary odds ratio estimate was 0.99 (95% CI 0.92, 1.06). In conclusion, our study does not support a causal effect of circulating uric acid on diabetes risk. Uric acid-lowering therapies may therefore not be beneficial in reducing diabetes risk.