A qualitative and quantitative characterisation of CD8+ T cell targets on human beta cells in Type 1 diabetes: perspectives from a proteomic strategy

© 2016 Dr. Kai Lin GiamType 1 diabetes (T1D) is a tissue specific autoimmune disease that result from the targeted destruction of the insulin-secreting beta cells in the pancreatic islets. A hallmark of T1D is the infiltration of autoimmune T cells into the islets that is predominantly composed of CD8+ T cells. The strongest genetic association with susceptibility to T1D are to the human leukocyte antigen (HLA) genes, particularly the class II, and to a lesser extent, the class I HLA genes. The class I HLA genes encode for cell surface HLA molecules that constitutively present endogenous antigenic peptides for scrutiny by incoming CD8+ T cells. Autoantigens involved in T1D have been widely characterised, however, little is known about the identities of autoantigenic peptides presented by less studied alleles and the nature of naturally processed and presented CD8+ T cell epitopes that leads to a targeted beta cell destruction as disease unfolds. In this study, the nature of bound peptide antigens (immunopeptidomes) restricted to a panel of T1D associated class I HLA allotypes (HLA-A*01:01, -A*02:01, -A*24:02, - B*08:01 and –B*18:01) were studied in depth using advanced mass spectrometry techniques. First I assessed the immunopeptidomes of C1R cell lines that were individually transfected with each HLA allotype. An extensive class I ligand dataset (up to 18 600 naturally presented peptides) was identified in this study which allowed the extension and clarification of previously reported binding motifs for these class I HLA molecules. We found no differences in the nature of peptide presentation by T1D disease predisposing and protective HLAs. Overall, features (subcellular localisation and biological functions) of source proteins for all five HLA allotypes were similar and we observed that peptide sampling was not affected by source protein masses and lengths. Next, we examined the features of naturally presented autoantigenic peptides by class I HLA allotypes on a panel of surrogate beta cells (class I HLA expressing C1R cells supertransfected with key beta cell autoantigens including PPI, prepro-IAPP, IGRP and ChgA). Within the immunopeptidome, novel PPI, prepro-IAPP and ChgA peptides, alongside previously reported epitopes were identified. Over half of these candidate epitopes were considered as non-canonical ligands, often present in nested sets of overlapping peptides of heterogeneous length with signal sequence derived peptides over-represented in my datasets. More importantly, preliminary tetramer staining experiments performed herein showed that a nested set of HLA-A2 restricted PPI signal sequence derived peptides encompassing the PPI15-24 epitope attracted different populations of CD8+ T cells from patients, highlighting a potential, and novel role of epitope length heterogeneity in T1D pathogenesis. Finally, the characterisation of CD8+ T cell targets was complemented with the first qualitative and quantitative analysis of the class I immunopeptidome of a HLA-A2(+) human islet beta cells. Apart from the identification of naturally processed HLA-I peptides derived from beta cell antigen such as islet cell autoantigen-1, I confirmed the natural presentation of two HLA-A2 restricted PPI epitopes (PPI15-24 and Ins B10-18), and identified two novel, overlapping ChgA peptides. Moreover, a novel HLA-A2 restricted, beta cell specific PTM version of PPI15-24 was also identified to be naturally presented in my study. To our knowledge, this account of a naturally presented PTM peptide represents the first class I HLA restricted PTM identified in T1D. Furthermore, quantitation of the presentation levels of PPI15-24 and Ins B10-18 revealed that PPI signal sequence region derived epitope is more abundantly represented in the beta cell immunopeptidome, suggesting a beta cell specific antigen processing machinery for secretory proteins. The beta cell immunopeptidome study was further complemented with the proteomic coverage of the islet from the same donor. My proteomic coverage represent the largest human islet dataset reported to date, and reveals a robust beta cell specific post translational modifications that are known to be induced by inflammation, including deamidation by TGase2, W→Kyn by IDO, and for the first time confirmed the local presence and generation of spliced peptide species composed of an Ins C-peptide fragment fused to the ChgA derived WE14 peptide. These PTM neoantigens represents a rich source for peptide sampling by both HLA-I and HLA-II molecules in the islets as disease progresses. In summary, this study has provided key insights into the nature of beta cell CD8+ T cell targets that is beginning to lead us to understand the molecular events that culminate in the beta cells being singled out for autoimmune destruction during the pathogenesis of T1D