Type 1 diabetes (T1D) is an autoimmune disease, which results from the destruction of insulin-producing beta cells by self-reactive CD8+ T cells. These CD8+ T cells recognize class I peptide-HLA (pHLA-I) complexes expressed on the surface of beta cells and our laboratory has shown that their frequency is similar in the blood of diabetic and healthy donors. These results suggest that a universal state of 'benign' autoimmunity exists, and question the mechanisms that drive T1D onset. The switch toward autoimmunity could rely on the beta cells themselves, which might become more visible to the immune system. Moreover, beta cells secrete not only insulin, but also extracellular vesicles, exosomes, intact insulin granules, and other proteins contained in the insulin granules. All these components can possibly be phagocytosed, processed, and presented by dendritic cells in the pancreatic lymph nodes, and thus triggering the activation of self-reactive T cells.

My project therefore aimed at investigating the visibility of the beta cell under different types of external stressors, by characterizing the peptides presented at their surface within the pHLA-I complexes (immunopeptidome) and by assessing the contribution of certain molecular actors to their formation. More specifically, I have studied the effect of Coxsackievirus B infection, which is one of the main environmental suspects in T1D, as well as the role of interferon-alpha, which is the earliest signature cytokine observed in children at risk of developing the disease and which is produced locally post-infection.

By infecting a beta cell line with different strains of CVB, we observed that the virus reduces the visibility of the beta cell and only few viral peptides are presented. We also demonstrated that the CD8+ T cell response to these sequences is limited and predominantly naive. In contrast, exposing beta cells to interferon-alpha greatly increases their visibility. The amount of pHLA complexes presented is higher than in unstimulated cells, peptides restricted to HLA-B alleles are enriched and peptide diversity is increased. Some neo-epitopes were found to result from alternative splicing at the RNA level, at the peptide level, or from post-translational modifications. The proteasome and immunoproteasome have been identified as key players in the formation of many of these sequences.

These results provide a comprehensive mapping of antigenic targets exposed by beta cells for recognition by autoimmune T cells. These antigens offer the possibility of tracking autoimmune and anti-CVB responses during the natural history of T1D.

Mots-clés

Type 1 diabetes, human leukocyte antigen, immunopeptidomics, autoimmunity, Coxsackievirus B, interferon alpha