A multilevel approach to the complexity of etiopathogenesis of Type 1 Diabetes: from genetic risk to the loss of beta-cell communication

Various factors come into play in the onset and progression of Type 1 Diabetes (T1D) and for this reason it is defined as a multifactorial disease. This feature is evident since the early disease stages, during which the intervention of various factors (both genetic and environmental) is necessary for the onset of the disease. Moreover, during the progression of the disease, a number of inter and intra cellular phenomena interact, resulting in the complete loss of pancreatic beta-cell capacity to produce and release insulin. One can easily realize that a multilevel approach, taking into account the network of interacting factors, is necessary for better understanding of T1D. This thesis aims at approaching the multifactorial nature of T1D in two different aspects: 1) Etiology: interactions between T1D susceptibility genes were examined separately and the joint risk of their interaction was estimated, using the Bayesian probabilistic approach. 2) Progression of the beta-cell damage: this was implemented in a virtual portion of diabetic pancreas, by evaluating the effects of apoptotic events induced by autoimmune responses on communication between beta-cells, and hence on the ability of the pancreas to produce insulin. The damage was studied by implementing a stochastic mathematical model that simulates the complex relationships that exist within a virtual cluster made up of pancreatic beta-cell. As concerns the first aim, our results confirmed that HLA-DR is the most relevant susceptibility gene compared to INS and PTPN22 and proved that the INS and PTPN22 genotypes marginally influence T1D risk in all HLA genotype risk categories. The absolute risk conferred by genes simultaneously carrying high, moderate or low risk HLA and risk genotypes at the other two loci, compared with non-risk-associated genotypes at all three loci, was 19.8%, 6.6% and 2.2%, in the family cohort and 11.5%, 1.7% and 0.1% in the case-control sample, respectively. The present work represents, to the best of our knowledge, the first study based on both case-control and familiar data sets, showing the joint effect of HLA, INS and PTPN22 in T1D in a Caucasian population with a heterogeneous age of T1D onset, generalizing previous findings regarding data sets consisting of patients and controls <15 years published in literature. Finally, this study shows that a feasible and accurate risk assessment can be performed by applying the BN method. Whit regard to the second aim, at 31% (normoglycaemia) and 69% (hyperglycaemia) of dead beta-cells the system appeared to be sufficiently robust biologically to maintain regular Ca2+ ions oscillations guaranteeing an effective insulin release. Simulations at 84%, 94% and 98% grades (severe hyperglycemia) showed that intracellular Calcium oscillations were absent. In such conditions insulin pulsatility is not expected to occur. Our results suggest that the islet tissue is biophysically robust enough to compensate high rates of beta-cell loss. These predictions can be experimentally tested 'in vitro' quantifying space and time electrophysiological dynamics of animal islets kept at different glucose gradients. The model indicates the necessity of maintaining glycaemia within physiological levels as soon as possible after diabetes onset in order to avoid a drastic interruption of Ca2+ pulsatility with the subsequent reduction of insulin release.