Session 8: Molecular pathways to beta cell death

Type 1 diabetes (T1D) develops from a complex “dialogue” that is established between invading immune cells, which release a variety of chemokines and cytokines and putative immunogenic signals released by injured or dying β cells, leading to local inflammation. This dialogue is shaped by the host genetic architecture, key clinical/demographic features, and environmental factors such as a precipitating viral infection or endogenous “danger signals”. Accumulating evidence suggests that the nature and tempo of this inflammatory response varies by disease stage, with cytokine signatures predominating during the initiation of T1D (when there is probably a major contribution by the innate immune response) that are distinct from than those present during more advanced stages of autoimmunity and β cell decline.
Interferon-alpha (IFN-α) is a Type 1 interferon that signals through the IFN-α receptor (IFNAR1) and the downstream tyrosine kinase effectors Janus kinase 1 (JAK1) and tyrosine kinase 2 (TYK2) via interactions with STAT proteins. Children with high genetic risk of T1D present a type I IFN-inducible transcriptional signature that precedes the development of autoantibodies, while islets from living donors with very recent onset T1D display a prominent type I IFN signature. Data from our own group has shown that IFNα activates downstream signaling pathways within the β cell that create a feed-forward cycle of inflammation and self-antigen presentation. On the other hand, IFNa also triggers “defense” signals in beta cells, such as PDL1 overexpression, that may protect – at least temporarily – these cells against the immune assault. These findings are of interest, because drugs that target the IFNα-receptor have shown efficacy in mouse models of T1D, whereas JAK and TYK2 inhibitors are currently being tested in Phase 2 and 3 clinical trials for other autoimmune diseases. Given this background, we hypothesize that IFNα-mediated signalling orchestrates early β cell and immune interactions during T1D initiation and is a pathway that can be targeted therapeutically to prevent the progression of T1D.These and other related findings will be discussed at the lecture.
Background literature for the lecture
Eizirik DL et al Nature Rev Endocrinol, 5: 219-226, 2009 Marroqui L et al. Diabetes, 64: 3808-3817, 2015 Op de Beeck A, Eizirik DL Nature Rev Endocrinol, 12:263-273, 2016 Marroqui L et al. Diabetologia, 60:656-667, 2017 Coomans de Brachene A et al. Diabetologia, 61:636–640, 2018 Colli M et al. EBiomedicine “in press”, 2018

Prior to the onset of type 1 diabetes, there is a progressive loss of self-tolerance. Continued accumulation of auto-antibodies that recognize beta cell derived proteins and the coincident or subsequent activation and activity of auto-reactive T cells lead to the destruction of pancreatic beta cells and loss of insulin secretion.  However, important questions remain regarding events that surround the initial loss of tolerance and the subsequent failure of regulatory mechanisms to arrest autoimmunity and preserve an adequate number of functional beta cells.  It is increasingly evident that various processes can lead to the generation of different classes of neo-epitopes, including enzymatic and non-enzymatic post-translational modification of self-proteins that increase the affinity of epitope presentation and/or recognition by CD4+ T cells. Emerging data suggest that conditions of ER stress promote neo-antigen formation in beta cells, giving rise to an increasing array of neo-epitopes which have the potential to activate pathogenic T cells that may not be effectively eliminated by negative selection. Indeed, such neo-epitope specific T cells are present at elevated frequencies in the peripheral blood of patients with established type 1 diabetes and can also be found among T cells isolated from the pancreatic draining lymph nodes and islet infiltrating T cells of cadaveric organ donors with established diabetes. Together, these observations support a paradigm in which neo-epitope generation leads to succssive waves of immune activation, initiating a feed forward loop that can amplify the antigenic repertoire towards pancreatic beta cell proteins. Our ongoing studies seek to determine the stage of the disease process during which various classes of neo-epitope specific T cell responses appear by examining samples from at-risk individuals and to establish whether neo-epitope responses associate with particular disease endotypes.

Immune responses to inflammation during checkpoint inhibition
Ana Louisa Perdigoto MD PhD, Daniel Burkhardt PhD, Smita Krishnaswamy PhD, Nancy Kirkles-Smith, Jordan Pober MD PhD, and Kevan C. Herold, MD
Autoimmune endocrinopathies are among the most common immune related adverse events following treatment of cancers with checkpoint inhibitors (CPIs). Beginning with the introduction of drugs that block the PD-1/PD-L1 axis, a number of reports have documented the occurrence of autoimmune diabetes. In a recent combined cohort of patients from UCSF and Yale the incidence of autoimmune diabetes was 0.9% and the median time to presentation was 20 weeks after initiation of the CPI treatment. Clinical features of this form of diabetes include a high frequency of insulin deficiency, presentation with ketoacidosis, and insulin dependence. About 40% have at least one positive autoantibody, and more than 75% of subjects are HLA-DR4. There are increased pancreatic exocrine enzymes in about 40% of subjects and we have found that the total pancreatic volume is reduced in the CPI treated subjects who develop diabetes compared to those who do not. The majority of cases have occurred with anti-PD-1/L1 treatment and consistent with this we found that PD-L1 as well as IDO, but not CD80 or CD86 are induced on  beta cells when they are cultured with IFNg with or without TNF. We have found PD-L1 expression in vivo on beta and non-beta cells from patients with chronic pancreatitis, autoimmune pancreatitis, and CPI induced diabetes. Unexpectedly the expression of PD-L1 on beta cells identifies cells that are most susceptible to killing either when cultured with peripheral blood immune cells or even to IFNg alone. Consistent with this FAS and TRAIL are increased in expression on the PD-L1+ beta cells. Our findings suggest that in response to inflammatory mediators to which beta cells are exposed, they express PD-L1, IDO, but also become susceptible to killing. PD-L1 expression may identify beta cells that are at risk for immune damage.