Abstract
Inflammation, including reactive oxygen species and inflammatory cytokines in tissues amplify various post-translational modifications of self-proteins. A number of post-translational modifications have been identified as autoimmune biomarkers in the initiation and progression of Type 1 diabetes. Here we show the citrullination of pancreatic glucokinase as a result of inflammation, triggering autoimmunity and affecting glucokinase biological functions. Glucokinase is expressed in hepatocytes to regulate glycogen synthesis, and in pancreatic beta cells as a glucose sensor to initiate glycolysis and insulin signaling. We identify autoantibodies and autoreactive CD4+ T cells to glucokinase epitopes in the circulation of Type 1 diabetes patients and NOD mice. Finally, citrullination alters glucokinase biologic activity and suppresses glucose-stimulated insulin secretion. Our study define glucokinase as a Type 1 diabetes biomarker, providing new insights of how inflammation drives post-translational modifications to create both neoautoantigens and affect beta cell metabolism.
Introduction
Type 1 diabetes (T1D) is characterized as an insulin-dependent glucose metabolic disorder arising from inflammation of the pancreatic islets and amplified by autoreactive autoimmune B and T lymphocyte responses. Local inflammatory cytokines and reactive oxygen amplify post-translational protein modifications (PTMs), including deamidation, oxidation, carbonylation and citrullination, all capable of compromising immune tolerance. In particular, citrullination has been identified in a variety of tissues and studied extensively as biomarkers of rheumatoid arthritis (RA). Anti-cyclic citrullinated peptide antibodies (anti-CCP) and T cells arise early in RA, correlate with disease severity, and are now routinely used for the diagnosis of RA. An emerging group of PTM biomarkers important in T1D have been identified including islet cell autoantigen 69 (ICA69), insulin, glutamic acid decarboxylase 65 (GAD65), islet antigen-2 (IA-2) and zinc transporter 8 (ZnT8). Citrullinated-78-kDa glucose-regulated protein (GRP78) and -GAD65 were found to elicit vigorous B and T cell autoimmune responses in both human T1D and NOD murine disease.
Glucose homeostasis in humans is highly regulated by the activity of glucokinase, catalyzing glucose phosphorylation, a first rate-limiting step of glycolysis in the liver and pancreas. More than 600 mutations of the human glucokinase gene have been described, resulting in reduced glucokinase enzyme activity in pancreatic β-cells and in hepatocytes. As previously described, patients with maturity onset diabetes of the young (MODY) is linked to specific mutations of the glucokinase gene, classified as MODY2. Clinical trials of glucokinase activators have been investigated in patients with type 2 diabetes (T2D), including piragliatin, MK-0941, AZD1656 and dorzagliatin. In contrast to T2D and monogenic diabetes, much less is known regarding the role of glucokinase dysfunction in autoimmune type 1 diabetes (T1D).
Citrullinated proteins are the product of the conversion of arginine to citrulline, catalyzed by peptidylarginine deiminases (PADs) in the presence of Ca++ . There are five PAD enzymes (PAD 1–4 and 6) and levels are often increased with tissue inflammation. Among these, PAD2 is widely expressed in different tissues such as brain, spinal cord, spleen, pancreas, bone marrow, skeletal muscle but not detected in liver, kidney and testis. PAD2 expression and activity is increased in synovial fluid from anti-CCP positive RA patients and positively correlated with disease activity. Relevant to T1D, transcriptomic and proteomic profiling demonstrates that PAD2 is highly expressed in prediabetic nonobese diabetes (NOD) islets. In this work, we demonstrate the presence of autoimmune B and T cells to glucokinase epitopes in patients with T1D. Moreover, we report that inflammation causes glucokinase to undergo citrullination almost exclusively in the pancreas, but not in the liver of NOD mice. Citrullination alters the enzyme kinetics of pancreatic glucokinase and PAD2/PAD4 inhibitor partially restores impaired insulin secretion mediated by proinflammatory cytokines. The work suggests that citrullination of glucokinase is a marker of beta cell dysfunction as well as an autoimmune biomarker of T1D and support a potential therapeutic strategy of inhibiting PAD enzymes to restore beta cell metabolic pathways.
Results
GK citrullination in inflamed NOD islets and inflammatory stressed-beta cells
We first performed immunohistochemistry staining and confocal microscope to examine if protein citrullination is increased in islets from the spontaneous NOD murine model of T1D. NOD mice develop hyperglycemia, insulitis, and lymphocyte infiltration of the pancreatic islets as one model of human T1D. Protein citrullination in islets is already observed in 3.5 week old NOD mice without infiltration of lymphocytes (Fig. 1a, b) and increases in citrullination staining intensity with increasing age of the mice (Fig. 1a). Increased citrulline staining coincided with insulitis, as confirmed by hematoxylin–eosin (H&E) staining and anti-CD3 staining from 16 week old NOD mice (Fig. 1b). Overall protein citrullination is not detected in liver and pancreas extracts from age matched C57Bl/6 mice or in NOD liver, but is significantly increased in NOD pancreas (Fig. 1c). As one band of the extract matched the molecular weight of GK, PAD treated human GK served as a citrulline staining control.
a Representative immunodetection of citrullinated proteins in NOD pancreas tissue sections at different ages of NOD mice (between 3.5 and 16 weeks of age). Brown staining demonstrates positive immunohistochemical reaction for citrulline with anti-citrulline antibody. b Representative H&E of NOD pancreas tissue sections at different ages (between 3.5 and 16 weeks of age) and anti-CD3 at 16 weeks of age. c, d Representative tissue extracted proteins, 50 µg per lane, from liver (L) or pancreas (P) of C57Bl/6 (B6) and NOD mice (16 weeks old) were separated on SDS-polyacrylamide gel electrophoresis (PAGE), and protein spots were stained by anti-modified citrulline detection kit (c) and anti-GK (d). PAD-treated recombinant human GK (PAD-rhGK) serves as the positive control for protein citrullination (c). e The tissue extracted proteins from NOD mice (16 weeks old) were immunoprecipitated using anti-GK and immunoblotted with anti-peptidyl citrulline and anti-GK as indicated.
Glucokinase is the primary glucose sensor since even small fluctuations of its enzyme activity alter the threshold of glucose-stimulated insulin secretion in pancreatic β-cells. In liver, the major role of glucokinase is to regulate the glycogen synthesis. As previously reported and confirmed here, GK levels are expressed in significantly higher levels relative to total protein in liver compared to pancreas from both C57Bl/6 and NOD mice (Fig. 1d). We next immunoprecipitated GK with specific antibody to confirm GK citrullination in NOD pancreas. As expected, we captured significantly more glucokinase from NOD liver extracts compared to pancreas from 16 week old NOD mice (Fig. 1e). In contrast, the citrullinated glucokinase signal is significantly increased in pancreas compared to liver from NOD mice (Fig. 1e). These data suggest that little overall GK citrullination occurs in the liver, while significant pancreatic GK citrullination arises in the same NOD mouse as disease progresses.
To assess whether inflammation triggers GK citrullination in insulin secreting β-cells, a cocktail of proinflammatory cytokines (rmIFNγ and rhIL-1β) were used to model the insulitis using rat insulinoma cell line, INS-1 cells, which displays the key characteristics of pancreatic β cells. The level of protein citrullination was increased in cytokine-treated INS-1 cells over time (12, 24 or 48 h) by immunofluorescence and by flow cytometry (Supplementary Fig. 1a, b, respectively). As expected, GK+/citrullination+ populations of INS-1 cells were also increased after cytokine treatment compared to untreated cells in a time dependent manner (Supplementary Fig. 1c). Finally, cytokine triggered glucokinase citrullination in INS-1 cells was confirmed by immunoprecipitation with anti-glucokinase and immunoblot with anti-peptidyl-citrulline (Supplementary Fig. 1d). In addition, we incubated islets from non-diabetic human organ donors with proinflammatory cytokines (IFNγ alone or IFNγ and TNFα) and assessed the induction of citrullinated GK. Our results showed that citrullinated modified-GK was significantly promoted in human islets treated with IFNγ alone (Supplementary Fig. 1e) compared to untreated human islets. Collectively, these results illustrate overall pancreatic citrullination and that citrullination of GK can be specifically amplified by inflammatory cytokines in tissue, islets, and cell based systems.
.jpg){kind=logo}
