The role of innate immunity and viral infections in the pathogenesis of type 1 diabetes

Type 1 diabetes (T1D) is a chronic autoimmune disease, characterized by a selective destruction of insulin-producing pancreatic beta cells in genetically susceptible individuals, leading to overt hyperglycemia and to a need for life-long exogenous insulin replacement therapy. Nowadays, many of our current concepts about T1D pathogenesis result from autopsy-based studies of human pancreas and from investigations using rodent models; however, in the few last years, in addition to improved research tools, organized efforts [e.g. Juvenile Diabetes Research Foundation (JDRF) Network for the Pancreatic Organ Donors with Diabetes (nPOD)] have been developed to obtain high quality pancreas specimens from organ donors to study processes underlying T1D development. The autoimmune destruction of beta cells has been assumed to occur in a cell-mediated manner, with effector CD4+ T-cells and cytotoxic CD8+ T-cells playing a key role; however, increasing attention is now being directed toward the participation of innate immunity and inflammation in T1D, and increasing evidence seems to support this hypothesis. In addition, it has been established that the autoimmune process starts long before the clinical presentation of the disease, with a highly variable duration (from several weeks to years) of such prodromal stage; in this pre-clinical phase, disease-specific, though non-pathogenic, autoantibodies against beta-cell autoantigens, appear in various combinations and titers and can be used to determine the risk for T1D development. The histopathologic hallmark of T1D is represented by "insulitis”, which, classically, describes the pancreatic islets infiltration by immune-competent cells (predominantly cytotoxic CD8+ T-cells, but also CD4+ T-cells, B cells, macrophages, NK cells and dendritic cells) and which represents the manifestation of the autoimmune attack against beta cells. In T1D, being a multifactorial disease, both genetic predisposition and environmental factors promote the triggering of autoimmune response against beta cells, and enteroviruses, in particular coxsackieviruses B-group (CVB), have received a major attention, in particular for their capacity to infect human pancreatic beta cells, even if we are still lacking the demonstration of a strong causal association between enteroviral infections and T1D. The particular tropism of CVBs to beta cells and the antiviral response could be modulated by the local expression of specific viral receptors such as hCAR (human Coxsackie- and Adenovirus Receptor) and viral intracellular sensors for the viral genome. The melanoma differentiation-associated gene 5 (MDA5) and the retinoic acid-inducible gene I (RIG-I) are well-established viral dsRNA sensors, that induce type I interferon (IFN-I) responses with the subsequent synthesis of a large number of proteins, many of which with antiviral activities, such as ISG15 (interferon-stimulated gene 15) protein. Due to stressors, like viruses, beta cells may produce, in response, chemokines (such as CCL2 and CXCL10), which subsequently attract both adaptive and innate immune cells to pancreatic islets, by binding to their specific receptors (CCR2 and CXCR3, respectively). Adaptive and innate immune cells (such as neutrophils and γδ T-lymphocytes), producing cytokines, may, in turn, induce inflammation and exacerbate the autoimmune reaction. Finally, beta cells may express on the surface stress-induced molecules, like the MICA protein (a ligand for NK and CD8 T-cells), contributing to attract immune cells to pancreatic islets. Aims of the present study were to analyze pancreatic expression of molecules involved in the interaction between viruses and target cells such as hCAR, MDA5, RIG-I and ISG15, as well as components of the innate immune system such as neutrophils, γδ T-lymphocytes, CCL2, CCR2, CXCL10 and MICA in human T1D. For this purpose, we analyzed formalin-fixed and paraffin embedded pancreatic sections obtained from four different human cohorts. In particular, we used pancreatic sections of: 4 T1D (disease duration: 1-8 years), 4 islet autoantibody-positive non-diabetic (AAb+) and 6 islet autoantibody-negative non-diabetic (CTR) multi-organ donors from nPOD cohort; 4 T1D transplanted patients (samples were taken at the time of biopsy from transplanted pancreas because of T1D recurrence; one sample was taken also when whole organ was explanted due to transplant failure) from nPOD-Tranplantation (nPOD-T) cohort; 5 new-onset T1D living donors undergoing pancreatic biopsy (taken 3-9 weeks after diagnosis of T1D) from the Norwegian Diabetes Virus Detection study (DiViD) cohort; 3 T1D (1 recent onset and 2 long standing-disease duration: 7 and 14 years, respectively) and 5 CTR multi-organ donors from Siena cohort. We performed immunohistochemical experiments to study the expression of CXCL10, myeloperoxidase (MPO) enzyme (to identify neutrophils), CCR2 and the presence of γδ T-cells in pancreatic tissue; double or triple immunofluorescence with confocal microscopy analysis for hCAR, MDA5, RIG-I, ISG15 CCL2, MICA and insulin and/or glucagon or somatostatin in order to identify islet cells subset(s) expressing the molecules of interest. CCR2+ cells, MPO+ cells, and γδ T-cells were quantified in the whole area of each pancreatic section employing Arcturus XT system and the ratio between CCR2+/MPO+/γδ T-cells and each pancreatic section’s area was calculated: data were expressed as CCR2+ or MPO+ or γδ T-cells/mm2. The colocalization analysis of RIG-I, MDA5, ISG15, hCAR, CCL2 and MICA with insulin, glucagon and somatostatin was performed by Leica Application Advance Fluorescence (LasAF) software. Differences between groups were assessed by Mann-Whitney U test and p <0,05 was considered as statistically significant. CXCL10 was found expressed within islets of T1D nPOD donors (suggestively in all insulin containing islets), with a strong expression signal; pancreata from AAb+ showed several CXCL10 positive islets with a weaker signal compared to T1D, while CTR showed low or no CXCL10 expression signal. MPO+ cells were detectable in pancreatic exocrine tissue of T1D, AAb+ and CTR donors from the nPOD cohort, at higher number in T1D and AAb+ compared to CTR donors (p:0,0095 and p:0,038 respectively). MPO+ cells were also found scattered in pancreatic exocrine tissue of T1D nPOD-T donors. In this cohort, the number of MPO+ cells was comparable to that of T1D and higher compared to AAb+ and CTR from nPOD cohort (p:0,0159 and p:0,0043 respectively). T1D donors from the DiViD cohort showed MPO+ cells scattered in pancreatic exocrine tissue but also within several isles. The number of MPO+ cells was comparable to that of T1D and higher compared to AAb+ and CTR from nPOD cohort (p:0,0162 and p:0,0007 respectively). CCR2 was found expressed in scattered cells within pancreatic exocrine tissue from all nPOD donors, but not within islets (except in one AAb+ donor, in which CCR2 positive cells were also detected within three islets) and the number of CCR2+ cells tended to be higher in T1D compared to AAb+ and CTR. γδ T-cells were found scattered in pancreatic exocrine tissue from all nPOD donors (in some cases few γδ T-cells were identified around and within islets) and their number tended to be higher in AAb+ compared to T1D and CTR. hCAR was found expressed in islets of pancreata from both T1D and CTR from Siena cohort, almost exclusively within beta cells (colocalization rate between hCAR and insulin: 82.0±12.0% in T1D and 75.2±9.1% in CTR); hCAR was found expressed also in islets of pancreata from T1D nPOD-T donors, preferentially within beta cells. MDA5 was found expressed in islets of T1D, AAb+ and CTR from nPOD cohort. In AAb+ and CTR organ donors, MDA5 was expressed preferentially within alpha cells (colocalization rate between MDA5 and insulin: 9,99±8,34% and 12,73±8,52%, respectively. Colocalization rate between MDA5 and glucagon: 17,99±6,42% and 32,87±14,20%, respectively); in T1D there was a tendency to be expressed preferentially within beta cells (statistical significance was achieved only in one case: colocalization rate between MDA5 and insulin was 7,81 ± 3,11%, and between MDA5 and glucagon was 0,32 ±0,18%. In the other two cases analyzed there were no insulin containing islets). A preferential expression of MDA5 within alpha cells was seen also in CTR from Siena cohort (colocalization rate between MDA5 and glucagon: 40.1±7.1%. Colocalization rate between MDA5 and insulin: <5%). RIG-I expression was observed mainly in islet delta cells, both in T1D and CTR from Siena cohort (colocalization rate between RIG-I and somatostatin: 47.8±21.6% in T1D and 51.6±21.2% in CTR. Colocalization rate between RIG-I and insulin and glucagon: <5%, both in T1D and CTR). A delta-cell specific expression for ISG15 was observed in T1D and CTR from Siena cohort (colocalization rate between ISG-15 and somatostatin: 48.3±11.5% in T1D and 45.77±7.33% in CTR). CCL2 was found expressed in islets of T1D, AAb+ and CTR from the nPOD cohort, almost exclusively within beta cells (colocalization rate between CCL2 and insulin: 79.94±24.94%; colocalization rate between CCL2 and glucagon: 5.18±0.57%); its expression was absent within insulin deficient islets from T1D. CCL2 was found expressed also in islets of T1D from the nPOD-T donors, almost exclusively within beta cells (colocalization rate between CCL2 and insulin: 76.97 ± 4.88%. Colocalization rate between CCL2 and glucagon: 8.19±0,36%). MICA was expressed in few insulin containing islets with inflammatory infiltrates, preferentially within beta cells, in pancreata from several T1D nPOD-T donors. We did not observe any expression of MICA in nPOD donors. The access to 4 international pancreas organ biobanks gave us the opportunity to study the contribution of the innate immune response to T1D pathogenesis and the potential involvement of viruses, with specific reference to the pancreatic expression of molecules participating in virus-target cell interactions. Our results showed that hCAR, MDA5, RIG-I and ISG15 were specifically expressed in pancreatic islets and not in exocrine tissue, suggesting that islet cells could be equipped to respond to viral RNA intermediates. In addition, and notably, the differential expression of viral receptor CAR and of viral sensors among different islet cell subsets may well explain the specific tropism of enteroviruses to beta-cells. In particular, the differential expression of viral sensors within pancreatic islet cells, with RIG-I and ISG15 expressed mainly at the level of delta cells, both in CTR and T1D donors, while MDA5 preferentially expressed within alpha cells in CTR and in beta cells only in T1D (in addition to hCAR expression preferentially within beta cells in CTR and T1D), suggest that each islet cell subset is equipped with a unique pattern of viral defense, dependent on the clinical scenario. With regard to the contribution of innate immunity into T1D, assuming that viruses (in particular CVBs) could trigger the primary insult to pancreatic islets, we can speculate that stressed beta cells may secrete chemokines (like CCL2, CXCL10) and express on the surface stress-induced molecules like MICA that, through specific receptors (like CCR2) can attract adaptive and innate immune cells (like neutrophils, γδ T-cells) creating a scenario in which beta cells, innate and adaptive immune systems are engaged in intrinsic conversations. In conclusion, our results support the hypothesis of a key role of the innate immune system and inflammation in T1D pathogenesis and in recurrent insulitis in T1D transplanted patient; moreover immune infiltration seems to affect the whole pancreas and not only pancreatic islets.