Search

Type 1 diabetes is a complex disease characterized by the lack of endogenous insulin secreted from the pancreatic β-cells. Although β-cell targeted autoimmune processes and β-cell dysfunction are known to occur in type 1 diabetes, a complete understanding of the cell-to-cell interactions that support pancreatic function is still lacking. To characterize the pancreatic endocrine compartment, we studied pancreata from healthy adult donors and investigated a single cell surface adhesion molecule, desmoglein-2 (DSG2). Genetically-modified mice lackingDsg2were examined for islet cell mass, insulin production, responses to glucose, susceptibility to a streptozotocin-induced mouse model of hyperglycaemia, and ability to cure diabetes in a syngeneic transplantation model. Herein, we have identified DSG2 as a previously unrecognized adhesion molecule that supports β-cells. Furthermore, we reveal thatDSG2is within the top 10 percent of all genes expressed by human pancreatic islets and is expressed by the insulin-producing β-cells but not the somatostatin-producing δ-cells. In aDsg2loss-of-function mice (Dsg2lo/lo), we observed a significant reduction in the number of pancreatic islets and islet size, and consequently, there was less total insulin content per islet cluster.Dsg2lo/lomice also exhibited a reduction in blood vessel barrier integrity, an increased incidence of streptozotocin-induced diabetes, and islets isolated fromDsg2lo/lomice were more susceptible to cytokine-induced β-cell apoptosis. Following transplantation into diabetic mice, islets isolated fromDsg2lo/lomice were less effective than their wildtype counterparts at curing diabetes. In vitro assays using the Beta-TC-6 murine β-cell line suggest that DSG2 supports the actin cytoskeleton as well as the release of cytokines and chemokines. Taken together, our study suggests that DSG2 is an under-appreciated regulator of β-cell function in pancreatic islets and that a better understanding of this adhesion molecule may provide new opportunities to combat type 1 diabetes.

Stimulator of IFN genes (STING) is a central hub for cytosolic nucleic acid sensing and its activation results in upregulation of type I IFN production in innate immune cells. A type I IFN gene signature seen before the onset of type 1 diabetes has been suggested as a driver of disease initiation both in humans and in the NOD mouse model. A possible source of type I IFN is through activation of the STING pathway. Recent studies suggest that STING also has antiproliferative and proapoptotic functions in T cells that are independent of IFN. To investigate whether STING is involved in autoimmune diabetes, we examined the impact of genetic deletion of STING in NOD mice. CRISPR/Cas9 gene editing was used to generate STING-deficient NOD mice. Quantitative real-time PCR was used to assess the level of type I IFN-regulated genes in islets from wild-type and STING-deficient NOD mice. The number of islet-specific glucose-6-phosphatase catalytic subunit-related protein (IGRP)206-214-specific CD8+ T cells was determined by magnetic bead-based MHC tetramer enrichment and flow cytometry. The incidence of spontaneous diabetes and diabetes after adoptive transfer of T cells was determined. STING deficiency partially attenuated the type I IFN gene signature in islets but did not suppress insulitis. STING-deficient NOD mice accumulated an increased number of IGRP206-214-specific CD8+ T cells (2878 ± 642 cells in NOD.STING−/− mice and 728.8 ± 196 cells in wild-type NOD mice) in peripheral lymphoid tissue, associated with a higher incidence of spontaneous diabetes (95.5% in NOD.STING−/− mice and 86.2% in wild-type NOD mice). Splenocytes from STING-deficient mice rapidly induced diabetes after adoptive transfer into irradiated NOD recipients (median survival 75 days for NOD recipients of NOD.STING−/− mouse splenocytes and 121 days for NOD recipients of NOD mouse splenocytes). Data suggest that sensing of endogenous nucleic acids through the STING pathway may be partially responsible for the type I IFN gene signature but not autoimmunity in NOD mice. Our results show that the STING pathway may play an unexpected intrinsic role in suppressing the number of diabetogenic T cells.

Aims/hypothesis We hypothesised that human beta cells are structurally and functional polarised with respect to the islet capillaries. We set out to test this using confocal microscopy to map the 3D spatial arrangement of key proteins and live-cell imaging to determine the distribution of insulin granule fusion around the cells. Methods Human pancreas samples were rapidly fixed and processed using the pancreatic slice technique, which maintains islet structure and architecture. Slices were stained using immunofluorescence for polarity markers (scribble, discs large [Dlg] and partitioning defective 3 homologue [Par3]) and presynaptic markers (liprin, Rab3-interacting protein [RIM2] and piccolo) and imaged using 3D confocal microscopy. Isolated human islets were dispersed and cultured on laminin-511-coated coverslips. Live 3D two-photon microscopy was used on cultured cells to image exocytic granule fusion events upon glucose stimulation. Results Assessment of the distribution of endocrine cells across human islets found that, despite distinct islet-to-islet complexity and variability, including multi-lobular islets, and intermixing of alpha and beta cells, there is still a striking enrichment of alpha cells at the islet mantle. Measures of cell position demonstrate that most beta cells contact islet capillaries. Subcellularly, beta cells consistently position polar determinants, such as Par3, Dlg and scribble, with a basal domain towards the capillaries and apical domain at the opposite face. The capillary interface/vascular face is enriched in presynaptic scaffold proteins, such as liprin, RIM2 and piccolo. Interestingly, enrichment of presynaptic scaffold proteins also occurs where the beta cells contact peri-islet capillaries, suggesting functional interactions. We also observed the same polarisation of synaptic scaffold proteins in islets from type 2 diabetic patients. Consistent with polarised function, isolated beta cells cultured onto laminin-coated coverslips target insulin granule fusion to the coverslip. Conclusions/interpretation Structural and functional polarisation is a defining feature of human pancreatic beta cells and plays an important role in the control of insulin secretion. Graphical abstract

Reduced insulin secretion results in hyperglycaemia and diabetes involving a complex aetiology that is yet to be fully elucidated. Genetic susceptibility is a key factor in beta cell dysfunction and hyperglycaemia but the responsible genes have not been defined. The Collaborative Cross (CC) is a recombinant inbred mouse panel with diverse genetic backgrounds allowing the identification of complex trait genes that are relevant to human diseases. The aim of this study was to identify and characterise genes associated with hyperglycaemia. Using an unbiased genome-wide association study, we examined random blood glucose and insulin sensitivity in 53 genetically unique mouse strains from the CC population. The influences of hyperglycaemia susceptibility quantitative trait loci (QTLs) were investigated by examining glucose tolerance, insulin secretion, pancreatic histology and gene expression in the susceptible mice. Expression of candidate genes and their association with insulin secretion were examined in human islets. Mechanisms underlying reduced insulin secretion were studied in MIN6 cells using RNA interference. Wide variations in blood glucose levels and the related metabolic traits (insulin sensitivity and body weight) were observed in the CC population. We showed that elevated blood glucose in the CC strains was not due to insulin resistance nor obesity but resulted from reduced insulin secretion. This insulin secretory defect was demonstrated to be independent of abnormalities in islet morphology, beta cell mass and pancreatic insulin content. Gene mapping identified the E2f8 (p = 2.19 × 10−15) and Dlg2 loci (p = 3.83 × 10−8) on chromosome 7 to be significantly associated with hyperglycaemia susceptibility. Fine mapping the implicated regions using congenic mice demonstrated that these two loci have independent effects on insulin secretion in vivo. Significantly, our results revealed that increased E2F8 and DLG2 gene expression are correlated with enhanced insulin secretory function in human islets. Furthermore, loss-of-function studies in MIN6 cells demonstrated that E2f8 is involved in insulin secretion through an ATP-sensitive K+ channel-dependent pathway, which leads to a 30% reduction in Abcc8 expression. Similarly, knockdown of Dlg2 gene expression resulted in impaired insulin secretion in response to glucose and non-glucose stimuli. Collectively, these findings suggest that E2F transcription factor 8 (E2F8) and discs large homologue 2 (DLG2) regulate insulin secretion. The CC resource enables the identification of E2f8 and Dlg2 as novel genes associated with hyperglycaemia due to reduced insulin secretion in pancreatic beta cells. Taken together, our results provide better understanding of the molecular control of insulin secretion and further support the use of the CC resource to identify novel genes relevant to human diseases.

Type 17 immune responses, typified by the production of the cytokines IL-17A and IL-17F, have been implicated in the development of type 1 diabetes in animal models and human patients, however the underlying pathogenic mechanisms have not been clearly elucidated. While previous studies show that IL-17A enhances inflammatory gene expression and cell death in mouse β-cells and human islets, the function of IL-17F in pancreatic β-cells is completely untested to date. Here we show that IL-17F exhibits potent pathogenic effects in mouse β-cell lines and islets. IL-17F signals via the IL-17RA and -RC subunits in β-cells and in combination with other inflammatory cytokines induces expression of chemokine transcripts, suppresses the expression of β-cell identity genes and impairs glucose stimulated insulin secretion. Further IL-17F induces cell death in primary mouse islets. This occurs via Jnk, p38 and NF-κB dependent induction of Nos2 and is completely ablated in the presence of an inducible nitric oxide synthase (iNOS) inhibitor. Together these data indicate that IL-17F possesses similar pathogenic activities to IL-17A in mouse β-cell lines and islets and is likely to be a type 17 associated pathogenic factor in type 1 diabetes.

In type 1 diabetes, maturation of activated autoreactive CD8+ T cells to fully armed effector cytotoxic T lymphocytes (CTL) occurs within the islet. At present the signals required for the maturation process are poorly defined. Cytokines could potentially provide the necessary “third signal” required to generate fully mature CTL capable of killing insulin-producing β-cells. To determine whether autoreactive CTL within islets respond to cytokines we generated non-obese diabetic (NOD) mice with a reporter for cytokine signalling. These mice express a reporter gene, hCD4, under the control of the endogenous regulatory elements for suppressor of cytokine signalling (SOCS)1, which is itself regulated by pro-inflammatory cytokines. In NOD mice, the hCD4 reporter was expressed in infiltrated islets and the expression level was positively correlated with the frequency of infiltrating CD45+ cells. SOCS1 reporter expression was induced in transferred β-cell-specific CD8+ 8.3T cells upon migration from pancreatic draining lymph nodes into islets. To determine which cytokines induced SOCS1 promoter activity in islets, we examined hCD4 reporter expression and CTL maturation in the absence of the cytokine receptors IFNAR1 or IL-21R. We show that IFNAR1 deficiency does not confer protection from diabetes in 8.3 TCR transgenic mice, nor is IFNAR1 signalling required for SOCS1 reporter upregulation or CTL maturation in islets. In contrast, IL-21R-deficient 8.3 mice have reduced diabetes incidence and reduced SOCS1 reporter activity in islet CTLs. However IL-21R deficiency did not affect islet CD8+ T cell proliferation or expression of granzyme B or IFNγ. Together these data indicate that autoreactive CD8+ T cells respond to IL-21 and not type I IFNs in the islets of NOD mice, but neither IFNAR1 nor IL-21R are required for islet intrinsic CTL maturation.

CD8+ T cells play a central role in beta-cell destruction in type 1 diabetes. CD8+ T cells use two main effector pathways to kill target cells, perforin plus granzymes and FAS ligand (FASL). We and others have established that in non-obese diabetic (NOD) mice, perforin is the dominant effector molecule by which autoreactive CD8+ T cells kill beta cells. However, blocking FASL pharmacologically was shown to protect NOD mice from diabetes, indicating that FASL may have some role. FASL can engage with its receptor FAS on target cells either as membrane bound or soluble FASL. It has been shown that membrane-bound FASL is required to stimulate FAS-induced apoptosis in target cells, whereas excessive soluble FASL can induce NF-κB-dependent gene expression and inflammation. Because islet inflammation is a feature of autoimmune diabetes, we tested whether soluble FASL could be important in disease pathogenesis independent of its cell death function. We generated NOD mice deficient in soluble FASL, while maintaining expression of membrane-bound FASL due to a mutation in the FASL sequence required for cleavage by metalloproteinase. NOD mice lacking soluble FASL had normal numbers of lymphocytes in their spleen and thymus. Soluble FASL deficient NOD mice had similar islet inflammation as wild-type NOD mice and were not protected from diabetes. Our data indicate that soluble FASL is not required in development of autoimmune diabetes.

Type 1 diabetes results from T cell-mediated destruction of pancreatic beta cells. The mechanisms of beta cell destruction in vivo, however, remain unclear. We aimed to test the relative roles of the main cell death pathways: apoptosis, necrosis and necroptosis, in beta cell death in the development of CD4(+) T cell-mediated autoimmune diabetes.We altered expression levels of critical cell death proteins in mouse islets and tested their ability to survive CD4(+) T cell-mediated attack using an in vivo graft model.Loss of the B cell leukaemia/lymphoma 2 (BCL-2) homology domain 3-only proteins BIM, PUMA or BID did not protect beta cells from this death. Overexpression of the anti-apoptotic protein BCL-2 or combined deficiency of the pro-apoptotic multi-BCL2 homology domain proteins BAX and BAK also failed to prevent beta cell destruction. Furthermore, loss of function of the death receptor Fas or its essential downstream signalling molecule Fas-associated death domain (FADD) in islets was also not protective. Using electron microscopy we observed that dying beta cells showed features of necrosis. However, islets deficient in receptor-interacting serine/threonine protein kinase 3 (RIPK3), a critical initiator of necroptosis, were still normally susceptible to CD4(+) T cell-mediated destruction. Remarkably, simultaneous inhibition of apoptosis and necroptosis by combining loss of RIPK3 and overexpression of BCL-2 in islets did not protect them against immune attack either.Collectively, our data indicate that beta cells die by necrosis in autoimmune diabetes and that the programmed cell death pathways apoptosis and necroptosis are both dispensable for this process.

Apoptosis of beta cells is a feature of both type 1 and type 2 diabetes as well as loss of islets after transplantation. In type 1 diabetes, beta cells are destroyed by immunological mechanisms. In type 2 diabetes abnormal levels of metabolic factors contribute to beta cell failure and subsequent apoptosis. Loss of beta cells after islet transplantation is due to many factors including the stress associated with islet isolation, primary graft non-function and allogeneic graft rejection. Irrespective of the exact mediators, highly conserved intracellular pathways of apoptosis are triggered. This review will outline the molecular mediators of beta cell apoptosis and the intracellular pathways activated.